Conceptual Design
and Cos^Study
Sulfur Oxide Removal From
Power Plant Stack Gas
SORPTION by LIMESTONE or LIME
DRY Process
3repared for the National Center for Air Pollution Control
1968
TENNESSEE VALLEY AUTHORITY
-------
Hot Air Duct
1-
..~.,:;~~::.:~~?~7~~~~:,
Coal
Pulverizer
:. :':': ',~ -: " : ',''', A : :,' " ::. : > ,",'" , ", : " >" ': ',",::: -:
-------
On February 16, 1967; the National Center
for Air Pollution Control (Public Health Ser-
vice, U. S. Department of Health, Education,
and Welfare) entered into a contract with the
Tennessee Valley Authority (TVA) for a series
of conceptual design and economic studies to
be carried out by TVA on processes for reduc-
tion of sulfur oxide emissions from power gen-
eration. The purpose is to evaluate objec-
tively and realistically the merits of differ-
ent methods under consideration for sulfur ox-
ide control, with a common and uniform basis
used for comparison.
Work has proceeded on three processes:
(1) limestone injection (dry process), (2) use
of limestone in a wet-scrubbing process, and
(3) ammonia scrubbing. The present report
covers the dry limestone injection method.
The work has been divided in TVA as fol-
lows.
Project Supervision
Applied Research Branch
Division of Chemical Development
Cost Estimates
Applied Research Branch
Division of Chemical Development
Design
Mechanical Design Branch
Division of Engineering Design
Design Branch
Division of Chemical Development
Plant Tests
Process Engineering Branch
Division of Chemical Development
Research Staff
Office of Power
PREFACE
Steam-Electric Generation Branch
Division of Power Production
Report Preparation
Applied Research Branch
Division of Chemical Development
Research Staff
Office of Power
Steam-Electric Generation Branch
Division of Power Production
A major part of the evaluation has been
the analysis of findings by other organiza-
tions who have worked on limestone injection.
These findings, which are, of course, the ba-
sis for the conceptual design, are used piece-
meal throughout the body of the report as
needed. In addition, to serve as a general
reference, the work of most of the organiza-
tions ~uoted is summarized and presented in
appendixes. The contributions of the follow-
ing organizations are acknowledged.
The Babcock and Wilcox Com~any
Bergbau-Forschung GmbH (Germany)
Central Research Institute of Electric
Power Industry (Japan)
Combustion Engineering, Inc. - Detroit
Edison Company
G. and W. H. Corson Incorporated
Land Institute for the Protection of Air
and Soil (Germany)
National Center for Air Pollution Control
(NCAPC )
Research-Cottrell, Inc.
Resources Research Institute (Japan)
Steinkohlen-Elektrizitat AG (Germany)
University of Stuttgart (Germany)
Wisconsin Electric Power; Inc.
iii
-------
Preface
. . . . . .
. . . .
. . . .
Swmnary . . . . . . . . . . . . .. v
Status of Process Development. . . .. vi
Study Assumptions . . . . . . .. vi
Process Equipment. . . . .. .... vii
Costs. . . . . . . . . . . vii
Research and Development Needed . viii
Introduction
Status of Limestone Sorption
(Dry Process) ...... . . . .
Lime s tone Type . . . . . . . . .
Calcining and Hydration. . . . .
Particle Size. . . . . . . . . . . . .
Point of Injection . . . . . . .
Injector Design. . . . . . . . . . . .
Mass Transfer. . . . . . .
Sulfur Dioxide Oxidation . . . .
Summary. . . . . . . . . .
Study Assumptions and Design Criteria. .
Plant Location . . . . . . . . .
Plant Size. . . . . . . . . . .
Fuel Type ..............
Sulfur Content of Coal . . . . .
Limestone Type . . . . . .
Calcination. . . . . . . . . . .
Amount of Limestone Injected. . . . .
Particle Size. . . . . . . . . .
Boiler Type . . . . . . . . . . .
Point and Method of Injection. . . . .
Dust Collection. . . . . . . . . . . .
Waste Disposal. . . . . . . . . . . .
Stream Pollution. . . . . . . . . . .
Control Function. . . . . . . . . . .
Process Equipment. . . . . . . . . . .
Major Alternatives. . . . . . . . . .
Equipment Description. . . . . . . . .
Effect of Limestone Injection on
Power Plant Operation
Combustion . . . . .
Slagging and Heat Transfer
Erosion. . . . . . . . . . . . . . . .
Air Heater Operation . . . . . .
Dust Collection. . . . . . . . . . . .
Dust Handling. . . . . . .
Dust Utilization and Disposal. .
Water Pollution . . . . . . . .
Stack Height . . . .
Economic Evaluation. . . . . . . . . .
Sorbent Cost. . . . . . . . . .
Inve stment . . . . . . . . . . . . . .
Operating Cost . . . . . .
iv
CONTENTS
iii
Research and Development Needed. . . . .
Sorbent Efficiency. . . . . . . . . .
Fine Grinding. . . . . . . . . . . . .
Method and Place of Injection. . . .
Slagging and Erosion . . . . . .
Increase in Retention Time. . . . . .
Turbulence Promotion. . . . . . . . .
Dust Collection. . . . . . . . . . . .
Use of Spent Sorbent . . . . . .
1
Conclusions and Recommendations.
. . . .
3
6
9
10
11
12
16
17
17
Appendix I: Limestone Sorption
Studies by Various Organizations. . . .
Tennessee Valley Authority (Muscle
Shoals, Alabama) . . . . . . . .
National Center for Air Pollution
Control. . . . . . . . . . . . . . .
Bergbau-Forschung (Essen, Germany). . .
Landesanstalt (Essen, Germany) ....
Resources Research Institute
(Kawaguchi-Saitama, Japan) . . . . . .
Steinkohlen-Elektrizitat AG (Essen,
Germany) . . . . . . . . . . . . . . .
Technical University of Stuttgart,
Institute of Process Technology
(Stuttgart, Germany) . . . . . . . . .
Central Research Institute of Electric
Power Industry (Tokyo, Japan) ....
Wisconsin Electric Power Company
(Milwaukee, Wisconsin) . . . . . . . .
Combustion Engineering and Detroit
Edison (Detroit, Michigan) . . . . . .
19
19
19
19
19
19
20
20
20
2l
21
23
23
23
23
24
24
26
Appendix II: Limestone Availability
and Technology. . . . . . . . . . . . .
Location and Nature of Deposits. . . .
Limestone Composition and Quality. . .
Commercial Production and Shipping
Grinding . . . . . . . . .
Calcination. . . . . . . .
Hydration. . . . . . . . . . . .
28
28
28
29
30
30
31
31
32
32
34
34
38
39
Appendix III: Experimental Work
Supplementing the Design Study. . . . .
Static Tests. . . . . . . . . . . . .
Injection Tests: Part Stoichiometric.
Injection Tests: Full Stoichiometric.
Appendix IV: Cost Estimates
Thermal Changes in Sorbent Injection:
Effect on Operating Cost. . . .
Cost and Benefit of Fine Grinding
Detailed Estimates. . . . . . . . . .
Appendix V:
Drawings.
. . . .
. . . . .
43
43
43
43
43
44
46
47
47
48
51
51
53
54
55
56
57
59
59
61
62
66
66
68
69
71
72
74
76
76
76
76
79
79
80
83
90
-------
SULFUR OXIDE REMOVAL FROM roWER PLANT STACK GAS:
CONCEPTUAL DESIGN AND COST S~JDY
SORPTION BY LIMESTONE OR LIME:
DRY PROCESS
The growing problem of atmospheric pollu-
tion by sulfur oxides in power plant stack
gases has generated a large amount of research
and development on removal methods. Of the
several processes that have been proposed, in-
jection of limestone or lime into the boiler
is the simplest and requires the least invest-
ment. Although the sulfur oxides are recover-
ed as calcium sulfate or sulfite, which are
likely to be useless products in most instan-
ces, the low investment makes the process at-
tractive. Moreover, the net cost penalty in
producing power can be predicted with fair ac-
curacy, whereas the net cost in recovery pro-
cesses is uncertain because the market for the
products is subject to variation.
Limestone injection also has the advan-
tage that it can be operated intermittently--
shut down when atmospheric conditions are such
as to give good stack gas dispersion--as a
means of reducing cost. Recovery processes,
on the other hand, must be operated continu-
ously for best economy.
Limestone or lime injected into the boil-
er reacts with the sulfur oxides and converts
them to calcium sulfate, which is removed
along with fly ash in the dust collecting
equipment. The reaction takes place only at
high temperature, however, so only a very
short time, on the order of a second, is avail-
able for sorption. Because of this, the sul-
fur oxides do not have time to diffuse very
far into the interior of the particles and
therefore only partial utilization of the lime-
stone is obtained. The degree of reaction can
be increased by putting a wet scrubber on the
end of the system; the limestone is all cal-
cined to quicklime in the boiler and that part
which did not sorb sulfur oxides in the boiler
reacts fully in the scrubber (to form calcium
sulfate and sulfite) because of the better con-
tact and longer retention time. However, the
gas is cooled in the scrubber and must be re-
heated to restore thermal lift.
Because of its attractive features, lime-
stone injection has been studied by numerous
organizations, particularly in Germany, Japan,
and the United States. In the United States,
the National Center for Air Pollution Control
SUMMARY
(Public Health Service) has taken the lead in
developing the process; several projects are
being sponsored by NCAPC in various research
organizations. As part of this effort, NCAPC
and TVA have entered into a joint program of
research and development on the dry process
(no scrubber at the end of the system), in
which the problem is being attacked on a broad
basis with the purpose of culminating the work
by large-scale field trials in TVA power
plants. In these tests the objective will be
to optimize the process, refine cost data, and
provide design data for using the process in
the power industry.
The various projects under way that bear
on this joint effort are as follows:
1. NCAPC: In-house fundamental studies on
reaction rate and effect of limestone type.
2. TVA: (a) Fundamental studies on calcine
porosity as affected by limestone crystal-
lography and mineralogy, (b) conceptual de-
sign and economic study, and (c) plant-
scale tests.
3. Other contractors: Various studies on
dispersed phase sorption, regeneration of
sorbent, contactor type, and spent sorbent
utilization.
The conceptual design and economic study
mentioned abovel is presented in this report.
The purpose is to develop the best design pos-
sible from existing data, estimate capital and
operating costs, and recommend further research
and development needed.
A contract for the NCAPC-TVA plant tests,
to be carried out in two TVA boilers, was en-
tered into on November 22, 1967. Purchase and
fabrication of equipment is under way. The
equipment type and test program will be based
on the present conceptual design study, with
incorporation of new information from other
current projects as it becomes available.
lOne of a series of design and cost studies
in which various sulfur oxide-removal pro-
cesses will be compared on a uniform basis.
Limestone injection will be divided into two
parts; this report covers the dry process
and a subsequent one will take up use of a
wet scrubber at the end of the system.
v
-------
Status of Process Development
Many parameters affect the sorption effi-
ciency of injected limestone; the more impor-
tant are (1) intrinsic properties of the lime-
stone that affect degree of porosity after
calcination, (2) particle size, (3) tempera-
ture at point of injection, (4) retention time
in the boiler, (5) uniformity of distribution
in the boiler, (6) relative motion between
limestone particle and the stack gas (affects
mass transfer), and (7) rate of sulfur oxide
or sulfite oxidation. The complicated rela-
tionships between these factors are being
studied by several investigators but adequate
data for predicting degree of sulfur oxide re-
moval under a given set of conditions are not
yet available. The tests made thus far indi-
cate that no more than 25 to 30% removal can
be obtained by injection of the theoretical
amount of limestone; with further study the
system can perhaps be optimized to a higher
degree of sorption.
Tests have shown that limestones vary
widely in their ability to sorb sulfur oxides
but efforts to correlate initial properties
with sorption efficiency have not yet been
successful. This problem is being studied by
NCAPC in its own laboratories and also by oth-
er organizations in the studies mentioned ear-
lier.
Particle size of the limestone injected
has been shown to be critical. On the basis
of present evidence, it is concluded that the
limestone should be ground as fine as is eco-
nomically feasible. Information is lacking,
however, on the cost of grinding to the very
fine size that appears desirable (on the order
of only a few microns in diameter), and on the
increase in limestone utilization that will re-
sult from the fine grinding. The data avail-
able on the latter were obtained in small-scale
tests and may not be applicable to large boil-
ers. A study of the cost of fine grinding is
planned by TVA under a NCAPC contract.
The temperature at point of injection is
a highly important variable. It must be high
in order to get fast reactions that will ac-
complish both calcination (decarbonation) and
sulfur oxide sorption in the very short time
available. But it must not be so high as to
"dead-burn" the lime and thus reduce its abil-
ity to take up sulfur oxides. The optimum
temperature appears to be somewhere between
2500° and 3000° F. This has not been estab-
lished conclusively, however, and the optimum
may well vary with limestone type.
It may be possible to get better results
by calcining the limestone (or calcining plus
hydration) before injection. This has been
indicated in small-scale tests but the results
are not conclusive. At the present state of
vi
development, it does not appear that the bene-
fit would offset the cost of calcination, but
further tests to clarify the point are planned
in the NCAPC-TVA plant-scale program.
Uniform distribution of the limestone in
the boiler is an obvious need but one that may
be difficult to accomplish. If the limestone
is ground to extremely small size, as seems es-
sential, it may be difficult to get good mix-
ing before the rapidly flowing gas carries it
away from the zone of reaction. Development
of an effective injection system is one of the
major objectives in the NCAPC-TVA tests.
The other major variables--limestone re-
tention time, gas-solid relative motion, and
oxidation rate--appear to be less important
and less work has been done on them. None of
the tests or proposals made so far show wuch
promise for increasing any of the three as a
means of improving sulfur oxide sorption.
Study Assumptions
The cost of limestone injection and, to
some extent, the type of equipment required,
depend on several factors such as sulfur con-
tent of the coal, size of installation, lime-
stone cost, degree of grinding, power plant
and quarry location, and excess limestone
(above the theoretical amount) used to increase
degree of sulfur oxide removal. It was neces-
sary to assume a set of conditions as a base
case for the conceptual design; in addition,
the effect of varying each parameter was ana-
lyzed and factors developed wherever possible
to relate the basic design and estimate to
other conditions. The basic conditions as-
sumed are given in the following tabulation.
Condi tions
Boiler size
Boiler type
Fuel type
Sulfur content of
coal
Plant location
Limestone type
Assumed for Study
200 megawatts
Pulverized fuel
Coal
3.5%
Northwestern Alabama
High-calcium limestone
(94.9% CaC03 + MgC03)
$2.05 per ton (deliv-
ered)
70% -200 mesh
200% of theoretical
8 feet above uppermost
burners
Approximately 2700° to
2800° F. (for lime-
stone only; injectors
tiltable to allow tem-
perature adjustment)
Limestone cost
Particle size
Amount of sorbent
Injection point
Injection temperature
-------
These conditions were selected as typical
or desirable on the basis of the best informa-
tion presently available. The most question-
able one is the particle size, which probably
must be smaller to get sorption efficiency up
to an acceptable level. However, the best
data for cost estimating were from tests in
which a particle size of this magnitude was
used. The cost of grinding to smaller parti-
cle size was estimated and is presented.
The amount of limestone assumed, 200% of
theoretical, is a compromise between low sul-
fur oxide removal and excessive solids load in
the boiler and dust-collection system; it is
questionable whether a solids loading higher
than this could be tolerated. At the present
level of development, this amount would do
well to remove 50% of the sulfur oxides. How-
ever, it is hoped that fine grinding and an
effective injection system will improve sorp-
tion efficiency.
Process Equipment
The equipment needed for limestone injec-
tion is relatively simple, which is the main
reason for the low investment. The main needs
are a limestone-receiving hopper, storage fa-
cilities, a grinding installation, an injec-
tion system, and conveyors between these units.
Both the storage and grinding systems will dif-
fer with size of plant; for example, a large)
multiunit power station could best be served
by a central grinding system with a single
large ball mill, whereas for smaller plants
roller mills may be preferable.
The injection equipment, which is the
most important part of the installation, rep-
resents only a small part of the investment.
All that is needed is an air blower or com-
pressor, transfer lines) and a set of nozzles.
These should be tiltable to allow variation in
trajectory--and hence variation in temperature-
time exposure of the limestone--which hopeful-
ly would adjust for differences in limestone
properties and in boiler operation.
Further capital cost is required for (1)
additional dust-collection equipment, to take
care of the increased solids load and possi-
ble adverse changes in ash resistivity, and
(2) additional ash-disposal equipment and pond
area. Tests have indicated that electrostatic
precipitator capacity should be increased by
about 200% for the twice-theoretical injection
assumed.
Costs
Capital requirements have been calculated
for several different conditions. The results
are summarized in Table 1; except for the items
noted) each figure is based on the conditions
assumed for the study (p. vi).
Table 1.
Capital Requirement
Capital, $/kw.
of installed
capacity
Exception to assumptions
(p. vi)
None
Calcined limestone vs.
uncalcined
Calcined and hydrated
stone vs. uncalcined
1000 vs. 200 mw.
7.75
17.25
lime-
20.15
3.95
The cost of the limestone is likely to
vary widely, depending on amount used) power
plant location, and shipping distance. In a
survey of the limestone industry made as part
of this study; it was indicated that the price
range will be roughly $1.00 to $2.00 per ton
(delivered, -1/2-in. size). If limestone in-
jection is adopted on a large scale, the prac-
tice may well develop of identifying the lime-
stone deposit best suited to a particular pow-
er plant (best combination of reactivity and
shipping cost) and setting up a quarrying op-
eration there if the deposit is not already
being mined. A very large amount of limestone
will be required if the process is operated on
a continuous basis--on the order of 700,000
tons per year for a 1000-megawatt plant (under
the assumed conditions). For intermittent op-
eration) the requirement would be about 63)000
tons for a total of 30 days' injection per
year.
Overall operating costs under various
conditions are listed in Table 2. These in-
clude capital cost charges (interest, depreci-
ation) and taxes) at 13% of "investment per
year.
Table 2.
Operating Cost
Exception to assumptions
(p. vi)
$/ton Mills/
of coal kw.-hr.
None
Grinding to 99% -325 mesh vs.
70% -200 mesh
Calcined limestone vs.
uncalcined
Calcined and hydrated limestone
vs. uncalcined
2.0% S in coal vs. 3.5%
Intermittent operation
(30 da./yr.) vs. full time
Limestone at $l.OO/ton vs.
$2.05
1000 vs. 200 mw.
1000 mw. plus $1.00 limestone
vs. 200 mw. plus $2.05
1.05
0.39
0.44
1.18
1.80
0.68
2.18
0.70
0.81
0.26
0.14
0.37
0.81
0.86
0.30
0.32
0.62
0.23
vii
-------
Research and Development Needed
Much remains to be learned about removal
of sulfur oxides by limestone injection. The
research and development needs can be divided
into three categories.
1. Effect of limestone type on sorption ef-
ficiency. This is a basic need; research
is proceeding under several NCAPC contracts.
2. Limestone utilization. The main oppor-
tunity for improving the limestone process
is to increase sorbent utilization. If it
is assumed that twice theoretical is the
largest amount that can be injected feasi-
bly, and that this would remove about )0%
of the sulfur oxides at the present level
of development, then an increase in utili-
zation would either improve removal effi-
ciency at the twice-theoretical injection
or would allow use of less lime if the )0%
removal were acceptable. It seems quite
unlikely that both could be attained, that
is, a high degree of removal with less than
twice-theoretical injection.
The main possibilities for improving
utilization--fine grinding, injecting at
optimum temperature, improving uniformity
of distribution, precalcination, and in-
creasing oxidation rate--are all being
viii
studied under various NCAPC contracts.
Others, including increase in limestone re-
tention time and in gas-solid relative mo-
tion, are under investigation but the tests
are being made in reactors quite different
from a standard power plant boiler. Tests
in standard equipment seem justified but
have not yet been planned.
3. Effect on power plant operation. There
are several problems in connection with
power plant operation that are supplemental
to the sulfur oxide sorption but are never-
theless quite important. These include
possible adverse effects of the limestone
injection on combustion and heat transfer
efficiency, corrosion, erosion, air heater
performance, dust collector efficiency,
dust handling, usability of ash, and water
pollution. Most of these are difficult to
study on a small scale; one of the major
sections of the NCAPC-TVA plant-test pro-
gram will be an extended run (probably 6
mo.) to identify long-term effects in re-
gard to these problems. In addition, dust
collector efficiency and ash utilization
are being studied under other NCAPC con-
tracts.
-------
SULFUR OXIDE REMOVAL FROM POWER PLANT STACK GAS:
CONCEPTUAL DESIGN AND COST STUDY
SORPTION BY LIMESTONE OR LIME:
DRY PROCESS
INTRODUCTION
The problem of pollution by sulfur oxides
emitted from power plants is rapidly becoming
a serious one. The amount emitted by power
plants burning fossil fuels in the United
States is at present equivalent to about 19
million annual tons of sulfuric acid and by
1970 will increase to about 24 million tons.l
Tall stacks have helped relieve the problem
but it is generally considered that reduction
of emission from many of the plants will be
required in the future.
Extensive work is being carried out on
processes for recovery of sulfur dioxide from
power plants in useful form.2 This is a com-
mendable objective for several reasons: (1)
the sulfur in coal or oil is a valuable na-
tional resource that is wasted in current
practice, (2) new commercial sources of sulfur
are needed because of the dwindling supply of
economically minable elemental sulfur (brim-
stone), (3) there is perhaps a chance that the
recovery process would give enough return from
sale of product to avoid any penalty on the
cost of power, and (4) the product would leave
the plant rather than accumulate and give rise
to a disposal problem. On the other hand, the
recovery methods that have been proposed are
all complex chemical processes that can hardly
fail to complicate operation of the power
plant. Moreover; the required investment is
high (on the order of $10-$30jkw. of capacity)
and operating costs are such that any failure
to sell the product or any major drop in the
price realized would increase power production
cost intolerably. Such an involvement in the
chemical industry. which might amount to as
much as 20% of the power plant iDcome, is not
attractive to most power producers. Other ad-
verse factors also enter, including difficulty
in finding a market for the very large quan-
tities of recovered product involved, continu-
ally decreasing load factor of the power plant,
1 Ludwig, J. H., and Spaite, P. W. Chem. Eng.
progr. 63 (6), 82-84 (1967); also, Rohrman,
F. A., and Ludwig, J. H. Chem. Eng. Pro gr.
61 (9), 59-63 (1965).
2 Slack, A. V. Chem. Eng. ~ (25), 188-196
(1967) .
and the problem of synchronizing the two unre-
lated operations. These disadvantages could
be reduced somewhat if a chemical company op-
erated the recovery unit as an independent en-
terprise, but even so there would be unavoid-
able complications in the operation and eco-
nomics of the power plant.
Alternatively, limestone or lime can be
used to sorb the sulfur oxides and the product
(calcium sulfate or calcium sUlfite) discarded.
This has several important advantages: (1)
there is no product that must be sold to avoid
heavy operating loss, (2) investment is rela-
tively low; (3) the process can be shut down
to save operating cost (if pollution regula-
tions allo~) when the weather is such as to
give good plum~ rise, and (4) the penalty on
power cost is predictable rather than variable
as in recovery processes (because of market
variations) .
These advantages have led to considerable
emphasis on limestone sorption, particularly
as an interim method to cover the period until
a recovery process is developed and proved.
In the United States the National Center for
Air Pollution Control (NCAPC) has devoted a
major part of its program to development of
the process, both by in-house research and by
contracts with various research organizations.
This program covers both the dry and wet
versions of limestone sorption. In the dry
method, limestone or lime is injected directly
into the boiler where sorption of sulfur ox-
ides takes place at high temperature. The re-
acted lime passes on through the system and is
removed with the fly ash in the dust-removal
equipment. In the wet system, the gas is
scrubbed with a lime slurry just before it en-
ters the stack; better utilization of the lime
is obtained but the gas must be reheated to
get the necessary plume rise.
The dry and wet methods each have advan-
tages and disadvantages. This report is re-
stricted to the dry process.
Development and proving of the dry pro-
cess has been undertaken by NCAPC and TVA as a
joint program, in which the work by all the
organizations involved will serve as the basis
for field trials in TVA power plants. The ob-
1
-------
jective is to move as rapidly as possible into
a full plant test program for optimizing the
process, refining cost data, checking long-
term effects on power plant operation, and
providing design data for use of the process
in the power industry. The various projects
under way or planned are as follows:
1. Small-scale.
a. Fundamental studies on reaction ki-
netics and variation in efficiency
among limestone types (NCAPC) TVA,
Illinois State Geological Survey) and
Battelle Memorial Institute). Work of
this type is also being done indepen-
dently by Bituminous Coal Research,
Inc., and the U. S. Bureau of Mines.
b. Use of spent sorbent as construc-
tion material (West Virginia Univer-
sity).
2. Pilot plant.
a. Injection tests (The Babcock and
Wilcox Company).
b. Fluidized-bed reactor (Esso Re-
search and Engineering Company).
c. Regeneration of spent sorbent (Esso
2
Research and Engineering Company).
d. Fluidized-bed boiler-reactor (Pope,
Evans and Robbins).
3. Conceptual design. Process design
study based on existing data, plus economic
evaluation of the process (TVA).
4. Plant scale.
a. Injection into an oil-fired boiler
(Florida Power and Light Company).
b. Injection into two types of coal-
fired boilers (TVA).
This report presents the results of the
conceptual design study on the dry process.
It is one in a series of design and cost
studies being made for NCAPC by TVA. The sec-
ond in the series will cover the wet process
for sorption by limestone or lime.
The present study will be used as the ba-
sis for the NCAPC-TVA plant test program, which
is being funded by NCAPC unaer a contract en-
tered into on November 22, 1967. Two TVA
boilers) of different types, are being fitted
with limestone-injection equipment by Babcock
and Wilcox under contract with TVA. Beginning
of actual tests is planned for mid-1969.
-------
STATUS OF LIMESTONE SORPTION (DRY PROCESS)
Limestone sorption by the dry process has
been studied by many investigators, in numerous
bench-scale and pilot plant studies and in a
few plant-scale tests. Notwithstanding this
bulk of work, so many chemical and physical
factors affect the process that the status is
not clear. In this first section of the re-
port, the work of the various investigators
will be analyzed as a background and basis for
the conceptual design.
Sorption by limestone is basically a sim-
ple method from the operating standpoint.
Finely ground limestone is injected into the
boiler at high temperature (in the range of
1500°-3000° F.), the particles sorb sulfur di-
oxide during their travel along with the gas
through the boiler, and the reacted particles
are removed from the gas stream in dust-collec-
tion equipment. No special reactor is re-
quired and no major changes in boiler opera-
ting procedure are necessary.
The dry limestone process has several se-
rious drawbacks, however, that make general
adoption questionable unless improvements can
be made. The main problem is incomplete reac-
tion of the limestone and sulfur dioxide during
passage through the boiler. For stoichiometric
injection of sorbent, based on the equation
CaO + S02 + 1/202
> CaS04'
investigators generally have obtained only 20
to 25% sulfur dioxide removal as a maximum.
Recycling may be a possibility but has not yet
been worked out. Therefore the major approach
to increasing removal has been to inject more
limestone, up to near three times stoichiomet-
ric in some tests. With such an excess, some
investigators have obtained 50 to 75% sulfur
oxide removal. Besides the obvious disadvan-
tage of raw material cost, such a large amount
of sorbent could well cause boiler operation
problems and would certainly require a large
outlay for additional dust-removal equipment.
The degree of sulfur oxide removal depends
on several interrelated chemical and physical
parameters, plus the fluid dynamics factors in-
volved in the relative flow of sorbent parti-
cles and gas. Only a beginning has been made
toward understanding the complex system that
results from the interplay of these factors.
The reactions that take place during sorp-
tion are complex and are not yet well under-
stood. The following are the ones usually as-
sumed.
1.
CaCOs .> CaO + CO2
This calcination step not only removes
the carbon dioxide to make room for the
sulfur dioxide but also develops porosity
as the carbon dioxide escapes from the par-
ticle interior. The resulting high surface
area is essential for rapid reaction of
sulfur dioxide.
A high temperature, 2000° F. or higher,
is required to make this reaction proceed
at an acceptable rate. The temperature can
be too high, however, in which case the par-
ticles sinter and lose porosity, a phenome-
non called "dead-burning." Time of exposure
is also important; a high temperature may
not damage the particle if exposure time is
short. 1
2. CaO + S02 + 1/202 ) CaS04
The oxygen for this reaction is supplied
by the excess air in the stack gas. Without
a sorbent such as calcined limestone, the
oxidation takes place only to a limited ex-
tent; in normal stack gas, only about 0.5
to 1.0% of the sulfur dioxide is oxidized
during passage through the boiler. With
limestone present, oxidation of the sulfur
dioxide sorbed is complete. Sulfite is ob-
tained only in reaction of Ca(OH)2 at low
temperature.
The optimum temperature for the reac-
tion is 1600° to 1800° F.
The sorption reactions are affected by
various parameters; these are listed and sum-
marized below and will be discussed in detail
in later sections of the report.
1. Calcitic Versus Dolomitic Limestone
The general term "limestone" covers a
broad range of materials'varying in calcium
and magnesium content--from calcite or ara-
gonite (CaCOs) to dolomite (MgCOs.CaCOs).
The terms "high-calcium limestone" (very
low in MgCOs) and "dolomitic limestone"
(approaching dolomite in MgCOs content) are
common in the trade. The MgCOs content af-
fects sulfur dioxide sorption because it
does not behave like CaCOs, either in re-
gard to calcination or reaction with sulfur
dioxide. In this report, the term lime-
stone is used in a general sense to cover
everything from calcite or aragonite to
dolomite.
2. Limestone Efficiency
Limestones vary widely in their chemical
composition and crystal structure, with the
result that material from one source may
differ from another in degree of porosity
developed during calcination, and, there-
fore, in sorption efficiency.
1 See Appendix II for further discussion of
calcination.
3
-------
3. Particle Size
As the retention time is quite short
and the sulfur dioxide molecules must dif-
fuse into the particle interior for high
sorbent utilization, small particle size is
highly important.
4. Precalcination
As a means of better utilizing the re-
tention time available in the boiler; the
limestone can be precalcined outside the
boiler. This adds greatly to capital and
operating costs, however, and there is some
question as to whether effective retention
time is actually increased, as will be dis-
cussed later.
Pretreatment can be carried a step fur-
ther by hydrating the calcined limestone.
A major advantage of this is that during
the hydration the lime breaks down to very
small particle size, on the order of a few
microns if the calcination and hydration are
properly carried out. Thus hydration is a
method of attaining very small particle size
without the expense of grinding.
5. Retention Time in Boiler
A sketch of a typical boiler is shown in
Figure 1. The stack gas flow begins at the
burners. From there it flows upward through
the main portion of the boiler, across the
convection pass, downward through the econo-
mizer (boiler water preheater) and air pre-
heater, through dust-collection equipment
(cyclone or electrostatic precipitator, or
both), and up the stack. Although this is a
fairly long path, the gas flow is so fast
(on the order of 50-75 ft./sec. in modern
boilers) that retention time in the entire
system is only a few seconds, in most cases
less than 5 seconds. Moreover, since the
temperature falls rapidly throughout the gas
flow path, only part of the retention time
is effective in sulfur dioxide sorption be-
cause in the latter part of the path the
temperature is too low. Thus in most boil-
ers the "effective" retention time is quite
limited, usually not more than 1 or 2 sec-
onds. Within this range there is consider-
able variation, depending on boiler design
and rate of firing.
6. Temperature
The temperature levels through a typical
boiler are shown in Figure 2. In the "fire
ball" in front of the burners, the tempera-
ture is 3000° F. or higher. As the gases
move upward from the burners they quickly
lose heat to the water in the tubes that
make up the boiler wall. In the convection
pass across the top, further heat is lost
to the superheater and reheater tube sec-
tions hanging in the gas stream. Finally,
the temperature is brought down to the final
4
level, usually about 300° F., by preheating
boiler water in the economizer and combus-
tion air in the air preheater.
Although this arrangement is fairly typ-
ical, there are many boiler designs and some
differ considerably from the one shown.
They are all similar, however, in regard to
the rapid decrease in temperature between
the 3000° and 1000° F. levels; below about
1000° F. the sulfur dioxide reaction rate
is so low as to be of little significance.
From the standpoint of retention time,
it is obviously desirable to inject the sor-
bent as early in the system as possible,
that is, with the coal. Although the tem-
perature at this point is not optimum for
sulfur dioxide sorption, because of the in-
creased dissociation pressure of calcium
sulfate, some sulfur dioxide will be sorbed
even at the high temperature. Whatever
sorption is accomplished during the drop to
optimum temperature obviously is a gain over
injecting at the optimum temperature, which,
as mentioned earlier, is 1600° to 1800° F.
Thus in actual operation in a boiler, the
sorption rate would start at a reduced level
depending on the injection temperature,
reach a maximum at 1600° to 1800° F., and
then decline as the temperature dropped dur-
ing travel through the convection pass.
Injection with the coal is questionable,
however; because of the danger of dead-
burning. There is some possibility that
raw limestone could be introduced at this
point; the situation is not clear, as will
be discussed. Calcined limestone or hy-
drated lime, however, has been shown to
dead-burn severely when injected at such
high temperature.
Hence pre calcined materials, and possi-
bly raw limestone also, are better injected
after the gas has cooled somewhat from
flame temperature. The optimum temperature
of injection, which is affected by several
factors, has not yet been established.
7. Sorbent Distribution
Because of the short retention time, it
is important that the sorbent be distribu-
ted rapidly over the boiler cross section
in order to contact all areas of the gas
stream with the proper amount of sorbent.
If much time is lost in effecting the prop-
er distribution, then some portions of the
stream will not have enough time for good
sorption after the sorbent becomes avail-
able.
Distribution is a difficult problem be-
cause of the very large cross section of
modern boilers (and the resulting large dis-
tance to the center), the rapid flow of the
gas, and the smallness of the particles.
-------
STACK
AIR
PREHEATER
\
DUST
COLLECTOR
..J
AIR
Figure 1.
STEAM SUPERHEATERS
AND REHEATERS
1 r-1 r""
I I I
, I, I
I I I
II I
I II I
! II .
I .' I
I " I
lJI I
l...J
'\
--
-- -
--
---
COAL
BURNERS
l
"'
,
,
,
,
,
'\
,
,-
/
,-
/
,
,
(
Typical Coal-Burning Power Boiler
In most injection systems proposed, the
particles are carried in a body of injec-
tion gas, so that the problem becomes one
of fluid dynamics--the mixing of one body
or jet of gas with another.
8. Mass Transfer
Since the particles are small, they
quickly assume cocurrent flow at about the
same velocity as the stack gas, which is not
conducive to good mass transfer from gas to
particle surface. There may be little that
can be done about this but several possibil-
ities have been advanced.
9. Rate of Oxidation
There is some evidence that oxidation
rate is important in the overall sorption
rate. If so, then catalysts, either added
or occurring naturally in the limestone,
may be helpful.
These are the major parameters that de-
termine the degree to which the limestone re-
acts with sulfur dioxide as it passes through
the boiler. In the following discussion, the
available data bearing on each of the factors
are evaluated and such conclusions as can be
made are presented. Also included are several
suggestions for process improvement, made by
various investigators, and the status of each.
As a basis for the discussion, typical
data from some of the test programs have been
assembled in Table I. These are presented
merely to illustrate the ranges of variables
that have been tested and the variation in re-
sults obtained; no general conclusions can be
drawn from the data because of the several vari-
ables involved. Reference may be made to Appen-
dix I for further data on sorption efficiency.
5
-------
POINT
I
2
:3
4
5
6
7
8
9
10
Ii
... -
I..... i I I I ...,
I I I I I I I I,
I I I I' Iii,
: I I i
'4' 5' ,71 ,811 I
"'-:Jo1 It: 6., ~ I 1'1 I
'--..I I I I I i I I
L...J I I I II ,
lJ I II I
lJ I I
l.J
TEMP..OF.
300
600
750
825
950
1200
1575
1780
1900
2100
>3000
20
AIR
PREHEATER
10
STACK
GAS
COMBUSTION
AIR
REHEAT SUPERHEATER
SECONDARY SUPERHEATER
45'
en
9
o
'\
CD
10
.
- -
- -
o::t
('IJ
- -
--
--
--
--
- -; ""'
II \
u
- BURNERS
/
BABCOCK a WILCOX
BOILER
....
"-
"
"-
...
"
"
"'
'\
~
/
"
"
/
"
/
"
(
Figure 2.
Temperatures in Typical Boiler
Limestone Type
Reports from several investigators show
that limestones vary widely in sorption effi-
ciency for sulfur dioxide. Hatfieldl (TVA)
found that reactivity could be related to the
porosity of the calcined material (raw lime-
stone has very little porosity), particularly
to the volume of the pores in the size range
of 17.5 to 0.035 microns; pores less than
0.035 micron in diameter did not appear to be
a factor in sorption. There was also good cor-
relation of sorption with density of the cal-
cine when pores larger than 17.5 microns in di-
ameter were excluded from the density determi-
nation.
l Hatfield, J. D. Report for Third Limestone
Symposium, Clearwater, Florida, December 4-8,
1967. See Appendix I.
6
These results were obtained from examina-
tion of samples of material calcined by NCAPC
for use in fixed-bed sorption studies. In a
report of these studies by Potter2 (NCAPC), it
was stated that the wide variation in sorptiv-
ity could not be explained by any difference
in composition. Kruel and JUntgen3 (Bergbau-
Forschung) have reported similar findings, as
has Dieh14 at Bituminous Coal Research (BCR).
The range of variation in sorption rate
is shown by results obtained by Van Heek and
JUntgen5 (Bergbau-Forschung). Under similar
2 Potter, A. E. Ibid. See Appendix I.
3 Kruel, M., and Jilntgen, H. 1£.i£. See Ap-
pendix 1-
4 Diehl, E. K. Ibid.
5 Van Heek, K. H., and Jilntgen,' H. IQg. See
Appendix I.
-------
Table I.
a
Removal of Sulfur Dioxide from Power Plant Stack Gas by Lime$one or Lime Injection
Percent
of Temperatuxe 802
stoichio- at injection removed,
Investi~tor Orp;anization Fuel Material injected Particle size metric point, of. %
TVAb Simulated gases Limestone 100% < 75}A 100 1425-1650 20
tI tI 200 " 37
400 72
600 90
DJlomite 200 23
400 48
TVAc 600 d 70
Coal Limestone 70% < 200 mesh 100 - 23
Wickert, K. Mobil Oil AG Oil Hy~rated calcined dolomite < 60.M 100 2400-2550 10-15
" " 100 40-45E
Kruel, M. Bergbau-FOrschung GmbH Coal 91% < 6oj-L 100 1650 20-45
" 200 60-65
300 80-;10
" 400 100
> 60.J-l 300 50
" < 6g,<.t 300 80
Impure dolomite 86% < 60)./ 200 25
" Limestone 41% < 6~ 200 65
c Same as coal 130 d 40-50
Pollock, W. A. Wisconsin Electric Coal Limestone
11 65 - d 0
Combustion Engineering c
and Detroit Edison Coal Iblomite Same as coal 100 - d 20
Zentgraf, K. M. STEAGc Coal Hydrated calcined dolomite 95% < 9O~ 200 2100 21-27
Calcium hydroxide 11 200 25-32
" Limestone 160 15-21
110 2730 20-30
Hydrated calcined dolomite 250 11 18-25
Goldschmidt, K. Stuttgart Universityc Coal Hydrated calcined dolomite 150 1830 20
11 270 " 32
Calcium hydroxide 150 41
11 280 53
Oil Hydrated calcined dolomite 150 33
11 270 50
Calcium hydroxide 150 51
" 270 72
Ishihara, Y. Central Research InstituteC Oil Calcium hydroxide 10;, 260 2100 62
Limestone 125 30
11 310 48
~ The test proceduxes used are described fuxther in Appendix I.
Small-scale tests.
c
d Plant-scale testE. assumed to be over 30000 F.
e Injected with the coal; temperatuxe
At one-third boiler load.
-:)
-------
conditions, the amount of sulfur dioxide sorbed
by three different limestones was 234, 160, and
95 normal cubic centimeters per gram. In tests
of the same materials by NCAPC (Harrington,
Borgwardt, and Potter),l the same general rela-
tion between activities of the three limestones
was found.
Diehl has reported finding only 15 "good"
limestones from a group of 85 from different
parts of the country. However, it should be
noted that calcining techniques were not uni-
form in these tests and that the calcination
conditions were not optimized for each partic-
ular limestone. It appears that calcine poros-
ity is governed both by intrinsic characteris-
tics of the limestone and by the calcining pro-
cedure (see Appendix II). Hence it is not
known to what extent the differences found in
the tests were due to variations in the lime-
stone.
The effect of impurities in the limestone
on calcine porosity is discussed in Appendix
II. The situation is not clear, but the main
indications are that sodium is helpful and sil-
ica harmful.
The relative efficiency of dolomitic and
high-calcium limestones is also important be-
cause in some areas only one of the two is
readily available. Dolomitic limestone, for
example, is rare in Kentucky but is abundant
in Minnesota and Wisconsin where there is lit-
tle high-calcium limestone.
NCAPC found no significant difference be-
tween dolomite and high-calcium limestone in
fixed-bed sorption rate tests. Ishihara2
(Central Research Institute), in pilot plant
tests, also found dolomite comparable with
other materials. Goldschmidt3 (Stuttgart Uni-
versity), however, obtained considerably bet-
ter results with hydrated lime than with h~-
drated dolomite calcine. Both Brocke's4 lLand
Institute for the Protection of Air and Soil
(Landesanstalt17 and TVA's (Appendix I) small-
scale tests also indicated inferior sorption
efficiency by dolomite.
Babcock and Wilcox5 (B and W), on the oth-
er hand, found that high-magnesium limestones
gave better sorption in pilot plant tests, an
l Harrington, R. E., Borgwardt, R.. H., and
Potter, A. E. "Reactivity of Selected Lime-
stones and Dolomites with Sulfur Dioxide."
Paper presented at the American Industrial
Hygiene Conference, Chicago, Illinois, May
1-5, 1967.
2 Ishihara, Y. Report for Third Limestone Sym-
posium, Clearwater, Florida, December 4-8,
1967. See Appendix I.
3 Goldschmidt, K. Ibid. See Appendix I.
4 Brocke, W. Ibid. See Appendix I.
5 Attig, R. C. Ibid. See Appendix I.
8
average of 21% removal versus 12%.
One of the factors not well understood is
the reactivity of the magnesium carbonate in
dolomite. Some have found no magnesium sulfate
in calcined dolomite exposed for a long period
to sulfur dioxide. Others, including Jilntgen
(Bergbau-Forschung) and Minnick6 (G. and W. H.
Corson), report reaction, although JHntgen
states that some samples react and some do not.
Hatfield found that dolomites are generally
more porous after calcination; hence, the ad-
vantage of higher porosity may offset low sorp-
tion by the magnesium oxide portion of the
dolomite.
Since the magnesium carbonate in dolomite
calcines more easily than the calcium carbonate,
inactivation by dead-burning is a consideration.
It has been shown, as will be discussed later,
that high-calcium limestone can be injected in
a high-temperature zone without dead-burning,
and that injection at this high temperature is
necessary to give good sorption in the limited
residence time available. Such a temperature
may be too high, however, for the magnesium
carbonate. In other words, to adequately cal-
cine the calcium carbonate in the very short
time available, it may be necessary to hard-
burn the magnesium carbonate. This is in line
with commercial lime-burning experience. 7
It is concluded that within each of the
two major limestone types--high calcium and do-
lomitic--there is wide variation in reactivity
among materials from various sources. Appar-
ently this is due mainly to variation in degree
of porosity developed during calcination, which
must result from differences in properties of
the raw limestone. The problem is the same as
in the lime industry, where high porosity is a
major objective in calcination. However, al-
though the lime industry has had the problem
for a long time, little progress seems to have
been made in relating initial properties to
porosity. 8
Current research by NCAPC, TVA, and oth-
ers should throw some light on the problem. A
concerted effort is being made to relate phys-
ical and chemical properties of the limestone
to degree of porosity developed during calci-
nation, taking into account that (1) the tem-
perature and residence time optimum for one
limestone may not be optimum for another, (2)
the combination of small particle size and rap-
id heating rate in limestone injection into
6 Minnick, L. J. Oral report at Third Lime-
stone Symposium, Clearwater, Florida, Decem-
ber 4-8, 1967.
7 Boynton, R. S. Chemistr and Technolo of
Lime and Limestone, 134, John Wiley (1966 .
8 See Appendix II. Also, Boynton, op. cit.,
132-164.
-------
boilers is quite different from commercial cal-
cination practice (or from small-scale calcina-
tion tests conducted so far), (3) the way in
which impurities are combined into compounds
may be more important than simple impurity con-
tent, and (4) the mineralogy and crystallogra-
phy of the limestone are likely to be important
considerations.
In regard to relative efficiency of high-
calcium and dolomitic limestanes, the conflic-
ting data make it difficult to reach a conclu-
sion. The burden of proof, however, seems to
be on dolomite because of experience in the
lime industry. Boynton, for example, states
that "even if the CaO is soft burned, the hard-
er-burned MgO component influences a denser
quicklime of lower reactivity than a comparably
calcined high calcium lime." In commercial
calcination, this effect can be minimized by
increasing retention time and decreasing tem-
perature, steps that calmot be taken in lime-
stone injection.
Equilibrium considerations are also
against the dolomitic types. Reidl points out
that to get the same sulfur dioxide equilibrium
concentration as for calcium oxide at 1800° F.)
the temperature of the magnesium oxide must be
more than 400° F. lower. And although the ki-
netics are not known, the lower temperature
must be a handicap to dolomite. He concludes
that high-calcium limestones should be superior.
This does not mean, however, that all do-
lomitic limestones are worse than all high-
calcium limestones. Variations between lime-
stones (and between dolomites) may well be more
significant than differences between the two
general types. Further information on this
point should be forthcoming from current work.
A method for evaluating raw limestones will
be badly needed if the process is adopted wide-
ly. For a given power plant the various poten-
tial sources of limestone will have different
purchase and shipping costs, and the differ-
ences cannot be evaluated properly without
some knowledge of the relative reactivities.
Calcining and Hydration
One possibility for improving sorbent ef-
ficiency is to calcine the limestone before in-
jection, thereby eliminating one of the steps
that otherwise takes up part of the limited re-
tention time in the boiler. The degree of ad-
vantage this might give is uncertain. For one
thing, it may not be possible to inject the
calcined lime as early in the boiler as would
be feasible for raw limestone. It seems logi-
l Reid, W. T. In Final Report, "Fundamental
Study of Sulfur Fixation by Lime and Magne-
sia," Battelle Memorial Institute, June 30,
1966.
cal that the raw limestone could be injected
at a higher temperature without dead-burning,
because the evolution of carbon dioxide should
keep the particle temperature below the gas
temperature. Therefore, if the maximum allow-
able injection temperatures for the two mate-
rials are sufficiently far apart, the limestone
might be calcined in the boiler during the pas-
sage between the two boiler temperature levels,
in which case the precalcined limestone would
not have any advantage in residence time.
Zentgraf2 at Steinkohlen-Elektrizitat AG (STEAG)
found this to be true in tests with hydrated
lime. At the optimum temperature for the hy-
drate (2100° F.), limestone was relatively in-
effective. However, when the limestone was in-
jected at 2730° F., results were slightly bet-
ter than with the hydrate at 2100° F.
On the other hand, the boiler may not be
an efficient calcining device. The lime indus-
try, over a long period, has developed refined
and efficient techniques for obtaining maximum
reactivity of calcined limestone (Appendix II).
It seems likely that limestone calcined outside
the boiler by modern methods would be a better
sorbent than a calcine produced in the boiler.
Zentgraf's results indicate that there was lit-
tle difference; however, he was using hydrated
lime rather than calcined limestone and param-
eters such as original source of the two mate-
rials and calcination technique in preparing
the hydrate may have affected the comparison.
Perhaps limestone calcined in such a way as to
give maximum porosity would have given better
results.
Very few data are available on injection
of calcined limestone. A few pilot plant tests
have been reported, but full data on them are
not available; Diehl, for example, found that
precalcined limestone was not as effective as
that calcined in the furnace. Jlintgen used ox-
ides but did not report a comparison with raw
limestone.
Zentgraf injected calcined limestone in
plant-scale tests with poor results, which he
attributed to dead-burning during calcining;
the material seems to have been a commercial
product with unknown history. Commercial mate-
rials obtained for TVA tests have also shown
evidence of dead-burning.
Calcine injection tests have been made at
TVA (part-stoichiometric injection; see Appen-
dix III). In tests at 1900° to 2100° F., the
calcined material was much more effective than
raw limestone, as shown by the weight ratio of
sulfate to calcium in the recovered solids (the
reacted lime could not be separated from the
2 Zentgraf, K. M. Report for Third Limestone
Symposium, Clearwater, Florida, December 4-8,
1967. See Appendix I.
9
-------
fly ash to give direct data). The average sul-
fate to calcium ratio for calcined limestone
(seven tests) was 1.20 as compared with 0.46 for
uncalcined materials (four tests). Conditions
were similar except that the calcined limestone
did not come from the same source as the raw
limestones.
The TV A results on calcined limestone are
in general agreement with Zentgraf's data on hy-
drated lime. In both cases, injection at a tem-
perature such as 21000 F., low enough to prevent
dead-burning of calcine or hydrate, gave better
results with precalcined material. The benefi-
cial effect of higher temperature on raw lime-
stone efficiency that Zentgraf found was not ad-
equately evaluated in the TVA tests. Limestone
was tested at higher temperature, but the next
convenient injection point was at the burner
level, where the temperature probably was about
30000 F. Results were no better than at lower
temperature, perhaps because mild "hard-burning"
offset the effect of longer retention time.
Hydration of the calcine introduces the
factor of particle size, since with careful con-
trol during the hydration process the lime can
be broken down to near-micron size. Several in-
vestigators have tested hydrated lime. Zent-
graf, as discussed earlier, found that hydrated
lime was not superior to raw limestone when
each was injected at its optimum temperature.
Ishihara obtained somewhat better results with
hydrate than with limestone (approximately 62
vs. 47% removal at 2.58 times stoichiometric)
when both were injected at 20000 to 22000 F.,
which agrees with Zentgraf's data. Ishihara
did not inject limestone at a higher tempera-
ture where it might have been more effective.
(He injected limestone with the fuel but this
introduces the factor of dead-burning. )
Again, this work indicates that raw lime-
stone is the best choice if it is injected at
the proper point. There may be some difficulty
in achieving this, as will be discussed later.
If it should develop that precalcination
is beneficial, it might be desirable to do the
calcining at the power plant. Techniques and
equipment appropriate for this are discussed in
Appendix II. However, it seems unlikely that
there could be enough improvement in reactivity
to offset the high cost of calcination (see.
p. 42). It should be possible to resolve
this question in the plant-scale test program.
Particle Size
Fundamental studies by several investiga-
tors have shown that at least two mechanisms
are important in sorption efficiency: (1) dif-
fusion through small pores into the limestone
particle, and (2) formation of a relatively im-
permeable calcium sulfate shell over the parti-
cle surface as the reaction proceeds. Reducing
10
the particle size helps in regard to both of
these.
In all of the applied work small particle
size has been beneficial. Examples are (1) the
pilot plant tests by Ishihara, in which sorp-
tion varied inversely with the one-fourth power
of the particle diameter, (2) pilot plant work
of JHntgen, who improved sorption from 25 to
43% by changing from plus 6o-micron to minus
6o-micron material, and (3) plant tests by
Zentgraf, in which about one-third more 70% mi-
nus 90-micron material was required to get the
same sorption as with 95% minus 90-micron mate-
rial. In part-stoichiometric injection of lime-
stone and calcine at TVA, minus 325-mesh parti-
cles separated from the product solids averaged
47% conversion to calcium sulfate (11 tests) as
compared with 18% for the remainder (about 50%
-200 mesh).
There is some question whether decrease in
particle size is as helpful in the extremely
fine size range as for larger sizes. Attig (B
and W) has reported preliminary work indicating
14 to 29 microns to be the optimum range.
JHntgen tried l-micron calcium carbonate and
found it no better than larger size. It is pos-
sible that agglomeration of the very fine parti-
cles prior to mixing with the furnace gas may
partially negate the benefits of fine grinding.
If so, standard agglomeration inhibitors used
in the lime industry might be helpful.
Zentgraf found that sorption was better
with small particles, but that it did not in-
crease proportionally to surface area. He at-
tributed this in part to agglomeration.
External mass transfer may also be a fac-
tor in efficiency of very fine particles (see
discussion on mass transfer, pp. 16-17).
The fine particles should have less slippage
in the gas stream than larger ones, and there-
fore less relative motion between solid and gas.
The resulting decrease in rate of mass transfer
may offset to some extent the beneficial ef-
fects noted above. Zentgraf advanced this as
one of the reasons for his inability to corre-
late reactivity with surface area.
It has been proposed that the limestone
used be one that decrepitates on heating; such
limestones are said to occur. Others have pos-
tulated that the rapid decarbonation of lime-
stone particles in the boiler explodes them
and thus gives fine particles anyway. This may
be partially true, but in TVA tests particles
as large as those injected have been found in
the product.
Range of particle size is probably an im-
portant factor, although no data have been re-
ported on it. For example, a mass of particles
uniformly of 90-micron size probably would be
better than one that Qnly averaged 90 microns.
Calcining the latter without dead-burning the
-------
finer particles would be difficult. This rais-
es the possibility that particle size uniformi-
ty could be improved by screening or some type
of special milling and sorption thereby in-
creased because of the resulting uniform calci-
nation of all particles. On the other hand,
there may be no point in fully calcining the
larger particles because only the outer portion
has time to react.
It is concluded that fine grinding is one
of the more promising ways to improve sorption
rate. The question is how fine to grind, both
from the standpoint of grinding cost and of
possible adverse effect on mass transfer at ex-
treme fineness. Fundamental work under way on
reaction kinetics should help in relating ad-
vantage to cost. In addition, fine grinding
will be tested in the plant-scale program.
Point of Injection
Much of the work in plant-scale injection
projects has been aimed at identifying the op-
timum injection point. From theoretical con-
siderations and small-scale work, it has been
established that 1600° to 1800° F. is the best
temperature range for isothermal sorption by
calcium oxide. The rapidly decreasing temper-
ature in the boiler and the rapid gas flow,
however, are complicating factors.
As noted earlier, it is desirable to in-
ject the sorbent as early as possible in order
to maximize retention; if it were not for the
possibility of dead-burning, it would obviously
be best to inject with the coal so as to uti-
lize the entire gas path through the boiler for
sorption. This may not be practical, however,
particularly for materials such as calcium ox-
ide and calcium hydroxide; Goldschmidt, for ex-
ample, found that calcium hydroxide injected
with the coal dead-burned so severely that low
sorption resulted (13-20% range at stoichiomet-
ric). Zentgraf, as has been noted, reported
2100° F. as the optimum for injecting calcium
hydroxide. Sorption fell off rapidly at higher
temperatures.
Hence the injection should be made at a
temperature that gives the best balance between
retention time and dead-burning. This is fair-
ly clear for calcium oxide and calcium hydrox-
ide but some question remains in regard to raw
limestone. Zentgraf did not inject limestone
at a temperature higher than about 2700° F. be-
cause he was working with a boiler designed to
remove most of the ash as molten slag and would
have lost much of the sorbent in the bottom
slag flow. Ishihara injected with the coal and
stated that the temperature was too high, al-
though it was in the range 2550° to 2920° F.
Detroit Edison did not inject with the coal but
through one of the upper burner ports; since
this put the limestone in the flame area the
temperature was quite high. Sorption decreased
when the burners and injector were tilted in a
converging direction, indicating a dead-burning
effect.
In contrast to these findings, Wisconsin
Electric obtained relatively good sorption from
injection with the coal, as did TVA in similar
tests. Unfortunately, in neither case was it
possible to inject beyond the burners for com-
parison.
It is concluded that limestone can be in-
jected at relatively high temperature without
excessive dead-burning, presumably because car-
bon dioxide evolution cools the particle during
its short exposure to the high temperature.
The main question is whether injection with the
coal gives too high a temperature and conse-
quently enough dead-burning to offset the ad-
vantages of injection at this point.l
It should be noted that injection with the
coal probably would not be feasible in a cy-
clone-type boiler because the sorbent likely
would become entrained with the slag and flow
out of the boiler. Pulverized-fuel boilers
do not present this problem.
Another point of probabJe importance is
that boilers differ widely in flame tempera-
ture, size of the flame "ball," and retention
time of particles in the flame zone. Hence in-
jection with the coal may be feasible in some
boiler types and not in others.
If injection with the coal is found inad-
visable, then the injection probably should be
just above, or perhaps in the upper part, of
the flame area. The reason for this is that
the gas cools rapidly to the 2700° F. level
that Zentgraf found to be better than lower
temperatures. And since Zentgraf did not es-
tablish an upper temperature limit, perhaps
the optimum is at even higher temperature.
It should be noted that optimum injection
temperature very likely is tied to particle
size, since decrease in size should increase
the danger of dead-burning.
The ideal system may be one in which the
limestone is injected at or near the top of
the flame area, through injectors that can be
tilted up or down. By this means it may be
possible to "tune" the injection to give the
optimum temperature for initial contact of
particles and gas. This may be helpful not on-
ly for finding the optimum temperature for a
given limestone but also for taking care of
difference in limestones and, to some extent,
variations in load. It is known that lime-
stones differ widely in resistance to dead-
burning; hence the upper temperature limit for
a given material may not be the same as for
l See page 43 for status of plans for further
tests on injection with the coal.
11
-------
another that might be used in the
In the plant-test program it
to fit the boilers with injection
or more levels to allow injection
temperatures.
same boiler.
is planned
points at two
at different
Injector Design
The problem of injector design may be a
major one because of difficulty in getting
quick and adequate distribution of sorbent over
the boiler cross section. This apparently has
not been a major problem in the plant-scale
tests carried out thus far, perhaps because the
boilers were small with relatively small cross
section and slow gas flow.
The most complete data on distribution
have been given by Zentgraf (Appendix I).
Three systems were used: (1) high-velocity
jets extending only a few centimeters through
the boiler wall, (2) pipes (water-cooled) ex-
tending various distances into the boiler and
with a nozzle at the end, and (3) alloy tubes
containing spaced holes (about 8 in. apart) ex-
tending all the way across the boiler. The
various types gave about the same distribution,
a ratio of 1:1.35 between minimum and maximum
concentration. This was for injection in the
upper reaches of the boiler; for injection
points nearer the burners there was some in-
crease in concentration near the boiler wall
(measured at a point in the convection pass),
which was said to occur because some of the
limestone traveling near the center of the
boiler had become occluded with molten ash and
stuck to the tubes.
It should be noted that the measurements
of distribution uniformity in these tests were
made at a point where the gas was relatively
cool and sulfur dioxide sorption slow; no data
are available on distribution in the zone of
high-temperature and high-sorption rate.
Other investigators have reported reason-
ably good distribution for jets set into the
boiler wall. All the boilers have been small,
however; Zentgraf's, for example, was about 20
feet across and had a gas velocity of 27 feet
per second. In contrast, one of the TVA boil-
ersl is 52 by III feet in cross section and has
a gas velocity of about 75 feet per second.
The problem of limestone distribution in the
larger boiler is likely to be much more severe.
In considering injector design, the first
question is whether or not the limestone can be
injected with the coal, in which case distri-
bution should not be a problem and no special
injection system would be needed. The boiler2
1 A 1300-megawatt unit scheduled for operation
in 1972.
2 This applies only to the pulverized-fuel
type.
12
is designed to get good heat distribution and
this should also give good distribution of fine
solids over the cross section in the combustion
area.
If a special injection system is necessary,
three principal methods can be considered: (1)
incorporation of sorbent into a large volume of
gas entering the boiler at relatively slow
speed (such as the tempering gas recirculated
in some boilers), (2) incorporation into a
small volume of fluid (air; combustion gas, or
water) blown into the boiler at high speed (to
get penetration before merging with the main gas
flow), and (3) mechanical injection of parti-
cles without any carrier fluid.
Method (1) is restricted by the fact that
introduction of a large body of gas into the
boiler must be taken into account in the boil-
er design. Normally the only large gas volumes
introduced are the combustion air, and for some
boilers, the recirculating and tempering gas.
Hence, for an existing boiler, injection would
be restricted to the combustion air (injection
with the coal) for a pulverized-fuel boiler and
to the tempering or recirculating gas for boil-
ers so equipped. Either could be undesirable,
one because the temperature might be so high as
to cause dead-burning and the other because the
tempering gas is introduced so high in the
boiler that temperature and retention time
could be too low for good sorption. Moreover;
both tempering and recirculating gas flows
vary so with boiler load that they may not be
practical for injection; this will be deter-
mined in the plant-scale program.
These limitations apply only to existing
boilers. For a new one, provision might be
made for injecting a large volume of recircu-
lated gas at the optimum point over a wide
load range.
Method (2) offers more flexibility be-
cause the amount of injection fluid would be
so small that it presumably would not upset
thermal or fluid flow conditions in the boiler.
It should be possible, therefore, to place the
injectors at any point. The problem is design
of the system--number and placement of injec-
tors, angle of injection, and type and velocity
of injection fluid--to get adequate distribu-
tion.
Method (3) has the advantage that no fluid
carrier is injected and therefore any adverse
effect of the fluid on boiler operation is elim-
inated. The problem is projecting the small
particles through the viscous, fast-flowing
gas far enough to get adequate distribution.
Calculations indicate that particles small
enough for good sorption (less than 325 mesh)
cannot be driven very far into the boiler even
at an injection velocity pf half the speed of
sound.
-------
If high-velocity injection in a carrier
fluid is adopted, a major question will be lo-
cation of injection points in the horizontal
plane. Obviously, the ideal system would be
to insert tubes of various lengths into the
boiler so that the entire cross section would
be served by judiciously spaced injection noz-
zles; each of Zentgraf's nozzles, for example,
served only a cross-sectional area of 32
square feet. Such a system would have the ad-
vantage that delivery of sorbent to a given
point could be varied if maldistribution of
sulfur dioxide in the boiler were found.
The main problem with this is that, as
noted earlier, the limestone probably should
be injected at 2700° F. or higher. At this
temperature, tubes extending almost halfway
across a large boiler would be quite difficult
to maintain even though water cooled. More-
over, slag accumulation OD the tubes or fall
of slag from the upper reaches of the boiler
might bend or break the tubes.
Alternately, the injectors could be mount-
ed in the boiler walls with little or no exten-
sion into the boiler. In this case there would
be much more dependence on carrier-fluid veloc-
ity and penetration than for nozzles spaced
over the boiler cross section. It probably
would be desirable to inject at an angle into
the main gas stream in order to get maximum
travel toward the boiler center with minimum
amount and velocity of carrier fluid. This
should also increase retention time, which
seems to be a very worthwhile objective. It
should be noted, however, that increase in re-
tention time at high temperature will change
the optimum injection point temperature. Al-
though a certain temperature, say 2800° F.,
might be optimum for horizontal injection, it
could well be too hot if a countercurrent move-
ment were imparted to the particles and the re-
tention time increased thereby by a fraction
of a second.
Hence it can be postulated that a wall ar-
ray of injectors should be located a few feet
above the top of the flame area, angled down-
ward, and designed for tilting. Then if an in-
crease in limestone calcination temperature
were desired, the injectors could be tilted
more steeply to drive the injection fluid car-
rying the limestone down closer, or perhaps in-
to, the flame. And if a lower temperature were
needed, the tilt would be changed to an angle
closer to horizontal, much as the burners are
operated in some tangentially fired boilers.
The difficulty with this is that a very
steep or a very shallow tilt might prevent
good distribution of sorbent. In this respect
spaced nozzles inside the boiler would be supe-
rior because good distribution could be obtain-
ed even though the nozzles were rotated to vary
the distance to which the limestone was driv-
en downward toward the high-temperature zone.
For either system the ideal arrangement
would be to set the injectors in fixed posi-
tion for the maximum retention time consistent
with good distribution (straight down for noz-
zles inside the boiler and at a calculated op-
timum angle for wall injectors), and then move
the injector array up and down in the boiler
until the best position in regard to limestone
exposure temperature was found. This would be
a difficult procedure to carry out.
Alternately, the injection velocity might
be varied as a means of varying the time-tem-
perature exposure of the limestone. This
should work well for the spaced nozzles inside
the boiler but could upset the distribution
pattern of a wall-mounted injector array.
Moreover, it may well be that maximum feasible
injection velocity will have to be used to get
even partial travel over the boiler cross sec-
tion.
If some exposure to the flame could be
tolerated, wall injectors might be placed
close enough to the flame area that the lime-
stone could be driven down into the upper lay-
ers of the circulating body of gas. There
might then be more flexibility in allowable in-
jector tilt because the main gas stream would
be doing the mixing.
For wall mounting, it would be desirable
to have injectors on both sides of the boiler
(spaced along the long dimension) in order to
minimize the projection distance required.
The type of carrier fluid is also a con-
sideration. Air has been used, for conve-
nience, in most of the large-scale tests. For
an operating installation, however, recircu-
lated combustion gas would have the advantage
of avoiding the heat loss in the air leaving
the system. Gas for recirculation could be
taken after the dust-removal system, com-
pressed, and piped to the limestone-injection
system. Residual dust probably would have to
be removed before compression, however; and
the cost of this would have to be balanced
against the heat loss from use of air.
Water has also been considered as the car-
rier and tests are being carried out in Ger-
many. One of the advantages is easier and
cheaper handling through the injection system.
No information is available on the hydrodynam-
ics of distribution.
Steam is another possibility. It is
available in a power plant at high pressure,
possibly might reduce dead-burning, could im-
prove precipitator performance, and might be
quite helpful in a sorbent recirculation pro-
cess (see pp. 45-46).
Even tentative conclusions as to optimum
injection method are difficult because of lack
13
-------
of information on the particle and fluid dy-
namics factors involved. It has not been con-
sidered advisable to go very deeply into this
in the present study because there is little
to report from the sulfur oxide work of others
and any analysis based on theoretical fluid dy-
namics would be a substantial study in itself.
However, even a cursory consideration of the
problem gives rise to several questions:
1. Can the limestone particles be driven
the 25 feet or so to the center of large
boilers before the gas temperature drops
too low?
The jet of injection air is deflected
immediately by the flow of boiler gas, and
the degree of jet penetration at a given
distance downstream depends on the injec-
tion velocity, jet diameter, and relative
gas densities. Three empirical equations
relating to this problem have been found.
Patrick: l
~ = -0.85(~ )0,38
do P \do
(1 )
Shandorov:2
~ = jJlW2(~)2'55
do fJ2v2 ~o
+ ~~ + ~~~~cotO<
(2 )
Ivanov: 3
~o = ~~)1.3Ct)3 + ~ coto<
where x = distance downstream (ft.)
y = travel across boiler (ft.)
P = velocity ratio of main stream to
jet (assuming equal densities)
do = initial jet diameter (ft.)
Pl = density of main stream
(lb./cu.ft. )
P2 = density of jet (lb./cu.ft.)
w = velocity of main stream (ft./sec.)
v = velocity of jet (ft./sec.)
~ = angle between main stream and jet
(3)
Rearrangement shows that these equations
are roughly equivalent. However, Patrick
does not include density or injection angle
terms and Shandorov did not test angles over
90°. Therefore Ivanov's equation (based on
angles from 60°-120°) was used to calculate
horizontal travel under the following as-
sumed conditions.
l Patrick, M. A. J. Inst. Fuel, p. 425 (1967).
2 Shandorov, G. S. Zh. Tekhn. Fiz 21; 1
(1957) .
3 Ivanov, Y. V. Sovetskoe Kotloturbostroenie,
p. 8 (1952).
14
Injection velocity
Main stream velocity
Weight ratio of injec-
tion air to limestone
Amount of limestone
540 feet per second
60 feet per second
Number of injectors
Jet density
2:13
Twice stoichiometric
for 200-megawatt
boiler (Case IV in
Appendix IV)
12
Equal to density of
air
For horizontal injection (0(= 90°), the fol-
lowing values were obtained.
y, ft.
x, ft.
5
10
25
12
95
1450
This indicates that, with the limited
amount of injection air available, the jet
would not travel very far into the boiler
until it had traveled a long distance down-
stream. Penetration could be increased by
using fewer injectors so as to increase the
diameter of each jet) but this might result
in maldistribution because of distance be-
tween jets. Or injection velocity might be
increased, but even the 540 feet per second
(half the speed of sound) may be difficult
to attain.
This does not take into account any pos-
sible effect of the limestone on effective
density of the jet; no treatment, either
theoretical or empirical, of a cloud of par-
ticles in a jet was found. If the bulk den-
sity of the jet were used in the Ivanov
equation, a lO-foot penetration would occur
in 6 feet of travel downstream rather than
95 feet. However, this is probably not a
tenable analysis of the situation. It seems
likely that the limestone particles would
have to be extremely small, on the order of
0.1 micron or smaller, before their density
could justifiably be averaged with that of
the air--the reasoning being that unless
their motion is governed by Brownian move-
ment, as would be that of a heavy gas mixed
with the air, their momentum would not fully
supplement that of the air molecules in
pushing the jet further out into the boiler.
Instead, their motion would be governed by
Stokes' or Newton's laws and they would
therefore slip within the jet as a result
of the centrifugal force of the turning jet)
thereby losing most of the push they other-
wise would have. There still should be
some force imparted by the particles that
would help push the jet outward and this
probably could be calculated; no attempt to
-------
do this has been made in the present study.
It is also conceivable that the parti-
cle momentum would be great enough to re-
sist the turning motion of the jet and the
particle could therefore have a flatter tra-
jectory and travel farther across the boil-
er. However, if particles of the size of
whiting (on the order of 10-micron diameter)
were used, it seems likely that they would
stay with the jet, and when the jet broke up
any remaining particle momentum in the hori-
zontal direction would be dissipated imme-
diately.
To check the penetration that might be
expected by small particles without benefit
of the jet, formulas given by Lapple and
Shepherdl were used.
S 0 ~fr;'[l - e-~]
S 0 ~'o;nP, 10 ~.ftYuuot + }-Newton'S law
where s = horizontal distance (ft.)
fJsD = particle dens ity (lb. / cu. ft. )
r = gas density (lb./cu.ft.)
D = particle diameter (ft.)
Uo = initial particle velocity (ft./
sec. )
II- = gas viscosity jjb./(ft. )(sec.17
t = time (sec.)
--Stokes' law
The Stokes' law equation applies to Reynolds'
numbers below 2 and the Newton's law to Rey-
nolds' numbers above 500. Both types apply
during the horizontal travel of a particle.
Neither takes into account the inhibiting
effect on horizontal particle velocity of
the vertical gas velocity when flow is in
the turbulent region. Equations to do this
are so tedious to solve that the effort did
not seem worthwhile for this study. The
simpler equations given above are conserva-
tive, however, because they give higher
values than would be obtained if the verti-
cal component were taken into account.
Solving the equations for a 10-micron
particle, 540 feet per second initial veloc-
ity, and 1 second, gives 6 feet horizontal
travel assuming turbulent conditions and
0.4 foot for streamline flow. Actual theo-
retical travel into stagnant gas is some-
where between, and this is reduced by the
effect of the vertical gas flow. Hence a
10-micron particle will travel much farther
into the boiler if carried within a jet.
2. Can the degree of penetration and resi-
l Lapple, C. E., and Shepherd, C. B.
Chem. is (5), 617 (1940).
Ind. Eng.
dence time be increased significantly by an-
gling the jet downward into the boiler?
To answer this, Ivanov's equation was
solved for an angle of 120° (downward injec-
tor tilt of 30°). The results are as fol-
lows (values in parentheses are those given
earlier for horizontal injection).
y, ft.
x, ft.
5
10
25
9 (12)
89 (95)
1435 (1450)
This indicates that angular injection does
not help much in penetration. However,
since the jet is driven some distance down-
ward before it reverses, there may be a
significant effect on retention time.
3. To what extent does turbulence in the
boiler gas stream affect mixing?
From discussions with boiler designers,
this appears to be a very difficult ques-
tion, affected both by boiler design and
operating variables. The only good source
of information probably is the experience
of the boiler manufacturers.
4. To what extent does uneven mass flow in
the boiler affect the situation?
If more gas passes through one particu-
lar area of the cross section than through
another, then more sorbent should be sup-
plied to that area. In other words, any
nonuniformity of m~ss gas flow should be
matched by a similar nonuniformity in sor-
bent distribution. Boiler designers have
advised, for example, that for front-fired
boilers most of the gas sweeps up the back
wall. This particular type of nonuniform
flow may be an advantage in sorbent distri-
bution, for injection penetration may not
have to be very deep to get the proper
amount of sorbent in contact with most of
the gas.
5. How soon after injection is good mixing
needed?
As noted earlier, the limestone should
be injected at relatively high temperature
to ensure calcination before the tempera-
ture range for optimum sulfur dioxide sorp-
tion is reached. There is some question as
to where significant sorption begins--con-
current with the beginning of calcination
or at a later stage, perhaps near comple-
tion, of calcination. If the latter, then
good sorbent distribution is not needed un-
til the particles have traveled some dis-
tance downstream. This might be particu-
larly helpful in boilers designed for in-
jection of recirculated tempering gas in
the upper part of the boiler; the limestone
could be injected at the best point for
15
-------
calcination and would then be subjected to
an effective mixing action when it reached
the turbulent level where the tempering gas
enters.
All this indicates that there are several un-
answered questions regarding sorbent distri-
bution in the boiler, and that the experience
of the boiler designer with the particular
boiler in question is of major importance in
designing an effective injection system. It
may well be that good distribution can be ob-
tained in some boilers but not in others. An-
swers to several of these questions may be ob-
tained in the design study being made by
Babcock and Wilcox for the NCAPC-TVA plant
tests.
Mass Transfer
For good sorption, the sulfur oxide mole-
cules must pass rapidly through the gas film
on the surface of the sorbent particle, through
the layer of reaction product, and into the
pores produced by calcination. Hence mass
transfer at one point or the other may be the
controlling mechanism.
The problem has been treated by Walling,
et al.,l who concluded that under typical boil-
er conditions sulfur dioxide removal by 74-mi-
cron particles (200 mesh) would be limited by
mass transfer considerations to about 33%, and
that unless the mass transfer can be improved
consideration of chemical kinetics is point-
less. Plant tests have indicated that this
33% limit is reasonable.
The main promise for improving mass trans-
fer is by reducing particle size, as discussed
earlier. Walling, et al., point out that their
analysis is not applicable to particles smaller
than 74 microns because of the limited data
available, and that extrapolation cannot be
made without further experimentation. Some
data of this type, although limited, were ob-
tained in the TVA tests supporting this study
(Appendix III). Minus 44-micron calcined
lime was well reacted (67-100%) during travel
through the boiler. (Uncalcined minus 44-mi-
cron particles did not do nearly so well, per-
haps because they were not injected early
enough for adequate calcination before reach-
ing the area of maximum sulfur dioxide sorp-
tion. )
One difficulty in using particles of ex-
tremely fine size is that gas-film diffusion
may become limiting because of low relative
velocity between particle and gas. Some pre-
l Walling, J. F., Cherry, Jr., R. H., and
Levy, A. In Final Report, "Fundamental
Study of Sulfur Fixation by Lime and Magne-
sia," Battelle Memorial Institute, June 30,
1966.
16
liminary consideration was given to this prob-
lem in the current study. Particle motion
equations based on Stokes' or Newton's laws
are not very helpful even though they are said2
to be applicable if all particle-velocity terms
are redefined as relative to the velocity of
the fluid. The difficulty is that the equa-
tions assume zero velocity at equilibrium,
that is, at infinite time and in the absence
of any force other than the original particle
momentum. Thus use of the equations leads to
zero relative motion between particle and gas
at equilibrium, whereas it is known from prac-
tice that in a moving gas stream there is a
"slip velocity" between gas and particle at
fully accelerated particle velocity. An em-
pirical equation of this, for horizontal ducts,
is given by Leva.3
Up " Ug~ - 1.41 DPO.3~~00.5J
where Ug = gas velocity (ft./sec.)
Up = particle velocity at equilib-
rium (n./sec.)
Dp = particle diameter (ft.)
Ps = solids density (lb./cu.ft.)
Using this equation and a gas velocity of
60 feet per second, the following values are
obtained for slip velocity.
Particle diameter;
microns
Slip velocity,
n./sec.
50 10.0
30 8.5
10 6.0
5 5.0
1 3.0
This should apply also to vertical travel as
the terminal falling velocity for particles
smaller than 50 microns is only a few inches
per second.
These values indicate that the slip ve-
locity is adequate for good mass transfer,
since Borgwardt (NCAPC) found in fixed-bed
tests that gas velocity had no effect on sorp-
tion until it dropped below 3 feet per second.
Unfortunately, however, the empirical equation
used above was derived from data on particles
355 microns and larger in diameter. It is not
known if the equation holds in the minus 50-
micron range. Moreover, the situation probably
is much more complicated than the above simple
treatment indicates; turbulence in the gas can
2 Lapple, C. E., and Shepherd, C. B.
Chem. 2,g (5), 61 7 (1940).
3 Leva, M. Fluidization, McGraw-Hill
Ind. Eng.
(1959) .
-------
have a major effect,l particularly microturbu-
lence in the vicinity of the particle surface.
The possibility of interposing barriers
in the flow path to induce turbulence and rel-
ative motion of gas and particles has been
considered (see p. 47). Assuming that the
structural material problem could be solved~
that the induced pressure drop did not serious-
ly affect power production economics, and that
slag accumulation did not become a problem,
such barriers might be helpful. There is not
enough information available at the present,
however, to be of value in a conceptual design
study.
Diffusion through the reaction product
layer and the lime particle pores is also an
important consideration. Jilntgen and others
have concluded that low permeability of the
calcium sulfate layer is the main cause of low
limestone utilization. Work on diffusion into
the particle is continuing in several labora-
tories, both in this country and abroad. From
the practical standpoint, the main hope of
avoiding such limitations is by reducing par-
ticle size, but, as has been pointed out, this
may be limited by increasing difficulty in
getting good distribution and by the decrease
in particle-gas relative velocity.
Sulfur Dioxide Oxidation
Since investigators generally have found
only sulfate in the sorption product (no sul-
fite), it appears that oxidation is part of
the operational sequence involved in sulfur
dioxide sorption. The mechanisms and rates of
oxidation have not been determined, however.
The equilibrium conversion of sulfur dioxide
to sulfur trioxide is not high at the tempera-
tures involved but reaction of sulfur trioxide
with calcium oxide, which probably is a fast
reaction, should promote the sulfur dioxide
oxidation by removing the reaction product.
Or calcium sulfite may form first and then ox-
idize. Also, it is not known whether oxida-
tion is a significant factor in reaction rate
as compared with gas-film diffusion, pore dif-
fusion, or chemical reaction between lime and
sulfur oxide. There are some indications,
however, that oxidation is significant. Both
Wickert and Wahnschaffe (Mobil Oil AG) found
that limestones with a relatively high content
of iron were superior sorbents. Ishihara's
pilot-scale work confirmed this; limestones
with an iron oxide content higher than 1% were
roughly 50% more effective than those below
0.5%. Moreover, in the Still process, which
involves sorption in calcium hydroxide at low
temperature, the presence of promoters is def-
l Torobin, L. B., and Gauvin, W. H.
Chern. Eng. 22, 189-200 (1960).
Can. J.
initely helpful.
Iron oxide or iron salts impregnated into
the limestone surface have also been tested.
Results were erratic. TVA found improved sorp-
tion in static tests but Ishihara and others
did not in injection studies.
The importance of oxidation rate is sup-
ported also by Bureau of Mines work, in which
increasing the amount of excess air improved
sorption rate and presence of nitrogen oxides
in the gas was found to be helpful.
All this indicates that it may be desir-
able to select a high-iron limestone for in-
jection or, alternatively, add an oxidation
promoter to the limestone. Support for this
is not conclusive enough, however, to warrant
incorporating it in the conceptual design.
Summary
This review of prior work shows the com-
plexity of limestone injection. A complicated
series of interrelated factors must be consid-
ered in evaluating the method and evolving a
conceptual design.
In evaluating the existing data, an at-
tempt has been made to weigh each item in
terms of its applicability to injection into
an operating boiler. In this respect, most of
the small-scale work falls short because the
experimental conditions were so different from
that to be expected in practice. Most of the
tests have been of the fixed-bed type, with
large particles, simulated gas, and relatively
slow calcination. In contrast, extremely fine
particles probably will be necessary in prac-
tice and calcination will be almost instanta-
neous. The various dispersed phase studies un-
der way should provide better data.
The pilot and plant-scale studies made so
far have been helpful, but, as is usually the
case in such studies, it has not been possible
to cover the several parameters to the degree
desirable.
The following conclusions summarize the
status of limestone injection as indicated by
the most applicable data. It should be noted
that these are tentative, subject to change as
the various studies develop new information.
1. The sorption efficiency of various lime-
stones and dolomites varies widely, to the
extent that some types may not be usable.
For each proposed installation, it may be
necessary to estimate cost of delivered sor-
bent from various points in terms of its
sorption efficiency.
2. The limestone should be used as is rath-
er than as the oxide or hydroxide. Plant
tests have indicated that sorption is about
the same if each is injected at the point
of optimum temperature. (However, as this
has not been fully established, the cost of
17
-------
using the oxide and hydroxide has been es-
timated also in this study.)
3. Particle size should be as small as
economically feasible.
4. The limestone should be injected at
relatively high temperature) possibly 2700°
F. or higher. Injection with the coal may
be feasible but has not been fully proved.
5. The preferred injection system (if ad-
dition with the coal is not feasible) is a
series of wall-mounted injectors operating
on air or recirculated gas injected at high
velocity) and with velocity) vertical place-
ment) and injection angle calculated for
18
the best combination of time-temperature ex-
posure) retention time) and distribution
uniformi ty .
If structural considerations would per-
mit) a series of nozzles spaced over the
cross section inside the boiler would be
desirable. By this means it should be pos-
sible to vary the time-temperature exposure
of the limestone--as a means of "tuning"
the system to a particular limestone--with-
out incurring nonuniform distribution. How-
ever) tubes extending into the boiler at
this point probably would not last very
long.
-------
STUDY ASSUMPTIONS AND DESIGN CRITERIA
The conceptual design developed in this
study is based on a single specific plant sit-
uation, with alternatives included wherever
necessary to generalize the design. Although
there is a wide range among power plants in
regard to the various chemical, physical, en-
gineering, and economic factors involved, the
specific design presented should be useful as
a base case for evaluating limestone injection
in other situations.
Plant Location
As a basis for development of cost data,
the TVA Colbert power plant in northwestern
Alabama (16 mi. west of Muscle Shoals) was se-
lected. This plant is located on the Tennessee
River, has an overall capacity of 1300 mega-
watts (one 500-mw. and four 20J-mw. units),
and has coal-burning boilers of the pulverized-
fuel type.
The Colbert location should serve ade-
quately as a general case. The main variables
in plant location are distance from limestone
supply and availability of space for reacted
limestone disposal. In these respects the
Colbert plant is well located; limestone is
available from operating quarries about 10
miles away, and disposal space is not a major
problem because the plant is located in a farm-
ing area where additional space is available
and probably could be purchased at reasonable
cost. The same situation appears to hold for
many of the power plants in the Uaited States.
The main exceptions are plants located in or
near urban centers, where quarry operation may
be feasible only some distance away and dis-
posal space may command a premium or even be
unavailable.
To cover plant locations not as favorably
located as the one selected, the effect of
freight cost on limestone delivered price was
reviewed; the results are presented in the
cost estimate section. Costs of disposal also
were evaluated.
Plant Size
A 200-megawatt unit size was selected as
the base case for the study From the chart
of power plant sizes in Figure 3, it can be
seen that this is a typical midrange boiler
size. For new plants, many of the units are
much larger, on the order of 400 up to as high
as 1300 megawatts. However, many of these are
nuclear fueled; moreover, if economical re-
covery methods that give a salable product are
developed, they will likely be applied mainly
to large boilers to improve economics. Hence
the smaller boilers that are already operating
appear to be the main candidates for limestone
injection.
In addition to the base case, the capital
and operating costs for injecting limestone
into a 1000-megawatt unit were estimated.
Fuel Type
CJal was selected as the fuel because it
is the prevalent type. Many plants burn high-
sulfur oil, however, and limestone injection
would be an appropriate method for sulfur di-
oxide control. The main difference is that
since oil produces very little ash, oil-burn-
ing plants have little in the way of fly ash-
collection equipment or ash-disposal facilities.
For lime injection these would have to be added,
at considerable cost.
The amount of added dust-collection and
ash-disposal capacity needed would be the same,
however, for either oil or coal firing; the
difference is that the new capacity would be
added on in coal-fired plants but would be new
for oil-fired units. This could well give rise
to a problem in finding room for electrostatic
precipitators and procuring land for disposal
ponds. There is no way of generalizing on this
problem; each plant situation would have to be
considered separately. Space for precipitators
probably could be found in most instances, and
if disposal area could not be obtained at the
plant it would be necessary to transport the
spent limestone to a suitable disposal point.
Sulfur Content of Coal
The coal deposits in use vary widely in
sulfur content, as shown in Table II. Two
levels, 2 and 3.5%, were selected for the study.
It seems likely that coals of relatively low
sulfur content will be consumed rapidly as the
pollution problem grows, so that the upper
level, or even higher; will become prevalent.
For the present, however, the two levels se-
lected appear to bracket most of the typical
coals in use.
Limestone Type
The limestone selected is from a quarry
near the Colbert plant. Price was the main
consideration; because of the very low cost at
the quarry, freight is a major part of the cost
delivered to the power plant. Reactivity was
checked and found to be acceptable, in an in-
termediate range.
The limestone is a high-calcium type; a
typical analysis is given in Table III. The
proved reserve from which it is being quarried
is large, estimated at 8 million tons of about
95% calcium plus magnesium carbonate material.
19
-------
- ~ LESS THAN 10 yEARS
DOVER 10 YEARS
INCLUDES NUCLEAR UNITS
.- SOURCE: POWER SYSTEM
STATEMENTS TO FPC FOR
1950 AND 1960
-
I-
-
~ fZZC
-
~
'7
- ~
/:
~ ~
~ ~
'W/1 ~
".".
..
a
z
«
CD
w
N
0000
~ ~IOO
xet
1-3
~g
00 ~ 75
I-
Z Oz:
:.::,)-
La.. ...J 50
O~
>~
I-
...J
m
if
«
<.) 0
mco
MEGAWATTS - m 0
-N
i I
00
o
N
..LI950..J L1960-' L- 1970----.J l 1980
Figure 3. Distribution of Thermal Power Plants in the United States
(Actual and Projected)
150
125
25
cnmlO
mml'-
-I'I')V
I I I
000
00
(\IV
YEAR
Such a reserve would last some 10 to 20 years
in supplying the needs of a 1300-megawatt plant
such as Colbert; other unexplored deposits in
the area probably would extend the supply in-
definitely.
In some other areas, dolomitic limestone
will be mJre economical, at least on the basis
of cost per ton (see Appendix II). As has been
noted, however, such material may not be as ef-
fective as the high-calcium type. No data are
available for estimating this quantitatively,
as a basis for comparing the effective cost
with that of high-calcium limestone located
farther away from the power plant.
Calcination
It is assumed for the basic case that the
raw limestone will be injected into the boiler,
since the evidence indicates that calcination
before injection does not give sufficient ad-
vantage to justify the cost. To cover the pos-
sibility that pre calcination will be found de-
sirable, the cost of using calcined limestone
20
AGE OF UNIT
mmcnm
mmmm
-1'4")101'-
I I I I
0000
000
NVW
mmmmo
mmmmo
-1'1')101'-(1)
I I i I 1000
00000000
oOOOONra
Nq-CDCD---
o
o
~o
00
00
CD-
and hydrated lime has also been estimated.
Amount of Limestone Injected
Treatment levels of 100 and 200% of
stoichiometric were used as the basis for the
cost estimates. It was assumed that higher
amounts would cause too much trouble and ex-
pense in boiler heat transfer, erosion, dust
collection, and waste disposal. Further data
on this will be obtained in the NCAPC-TVA
plant test program.
Particle Size
Estimates were made, based on reported
experience in grinding limestone, for the cost
of grinding to various levels of particle
size--over the range from 70% through 200 mesh
(similar to the grind obtained in power plant
coal mills) to an average of 10 microns (about
99% -325 mesh). This was supplemented by an
estimate based on a semitheoretical relation-
ship used in the grinding industry. For the
basic injection cost estimate, 70% through 200
-------
Table II.
Sulfur Content of Coal from
Various Sourcesa
Location of mine
(county)
Pyritic Organic Sulfate Total
Clearfield) Fa.
Franklyn) Ill.
Union) Ky.
Webster) Ky.
Boone) W. Va.
Walker) Ala.
Jefferson) Ala.
Clay) Ind.
Lee) N. C.
Butler; Pa.
Allegheny) Fa.
McDowell) W. Va.
Letcher) Ky.
Morgan) Tenn.
Cherokee) Kans.
2.82
1.50
1.05
0.70
1.47
0.81
0.97
2.13
1. 52
0.47
0.79
0.08
0.13
1. 75
1.99
0.74
1.02
2.23
0.78
1.01
0.81
0.72
1. 79
0.80
0.62
0.66
0.46
0.51
1. 78
0.71
o
o
o
o
o
o
0.03
o
o
0.07
0.23
0.01
0.04
0.71
0.32
3.56
2.52
3.28
1.48
2.48
1.62
1. 72
3.92
2.32
1.16
1.68
0.55
0.68
4.24
3.02
a Field) J. H.) Brunn) L. W.) Haynes) W. P'7
and Benson) H. E. Bureau of Mines Report of
Investigations 5469) p. 3) U. S. Department
of the Interior (1959).
Table III. Typical Analysis and Physical
Properties of Colbert County Limestone
Chemical analysis, % by wt.
CaC03
MgC03
Si02
Fe 203
A1203
91.8
3.1
3.5
0.3
0.4
Physical properties
Color
Bulk density) lb. leu. ft.
Hardness (Hardgrove Index)
Specific heat) B.t.u./lb. at
1000 F.
Gray
150
50
0.205
mesh was assumed. An effort was also made)
based on pilot plant results by Ishihara and
Tanakal (Resources Research Institute) on ef-
fect of particle size) to evaluate the econom-
ics of fine grinding versus increase in amount
of limestone used. Confirming data on a full-
plant scale are needed) however) to allow a
dependable cost correlation between fine grind-
ing and amount of sorbent injected.
~ Tanaka) K. Report for Third Limestone Sym-
posium) Clearwater) Florida) December 4-8)
1967. See Appendix I.
Boiler Type
The boiler on which the study is based--
one of the 200-megawatt units at the TVA
Colbert Steam Plant (Fig. 4)--is of the pulver-
ized-fuel type, that is) the coal is finely
ground and injected through burners into the
boiler. This is the most common type. The
second most common) the cyclone type) differs
in that the coal is not finely ground and is
burned in cyclones (furnaces that discharge
heated gas into the boiler proper through
ports) set into the side of the boiler. Most
of the ash in the cyclone type leaves the
boiler as bottom slag) whereas most of it
leaves as fly ash in the pulverized-fuel type.
Thus there would be a lower total load of
solids in the gas from the cyclone boiler and
a much higher ratio of limestone to fly ash)
which would probably introduce differences in
slag formation on boiler surfaces and in re-
moval of solids from the gas leaving the sys-
tem. There is not enough information to eval-
uate these differences on the basis of cost.
For the boiler selected) the height is
120 feet and the cross section at the burner
level 24 by 58 feet. Gas flow rate in the
main portion of the boiler is about 60 feet
per second) which gives a retention time of
1.5 seconds between a point just above the
burners and the 12000 F. point in the convec-
tion pass. The boiler is of the balanced-
draft type (fitted with both forced-draft and
induced-draft fans).
Point and Method of Injection
The base case is injection about 8 feet~
above the burners through six high-velocity)
tilting nozzles set into the boiler wall at
one elevation and spaced across the length of
the firing face of the boiler. Air is used as
the injection fluid since adequate data are not
available on cost and practicality of using re-
circulated gas. Also) injection pipes extend-
ing into the boiler are not considered because
information is not available on practicality
of maintaining them in the boiler.
More injectors may be needed for adequate
distribution; again) adequate data on the fluid
dynamics of mixing have not been developed.
The required amount and velocity of injection
air also are not known. Adequate compressor
capacity has been included to take care of a
wide range in these factors. A minimum of air
should be used) of course) in order to reduce
adverse thermal effects in the boiler) pressure
drop across the convection pass) and dry gas
2 This is only a rough estimate and must be
refined by the fluid dynamics analysis to be
made in conjunction with the NCAPC-TVA plant-
scale tests.
21
-------
- - - -
£/..-£,l':L Q~...DRU1'll -
?
1
1
,"-
"-
-0
...
.,-:
rJ
d
Figure 4.
200-Megawatt Unit at TVA Colbert Steam Plant
22
-------
heat loss from the stack; further information
is needed on the minimum that will give ade-
quate sorbent distribution.
Dust Collection
Some power plants have only cyclone dust
collectors) some have electrostatic precipita-
tors) and some have both. It is assumed for
the base case that dust will be removed by
electrostatic precipitators) on the basis that)
because dust is also a major pollution problem)
plants not now equipped with such equipment
will be by the time sulfur dioxide removal is
practiced. It is also assumed that dust emis-
sion must not be increased because of the lime-
stone injection.
Since the limestone injected will increase
the dust loading) the cost of additional pre-
cipitator capacity required to prevent dust
emission from increasing has been estimated.
For stoichiometric injection) it is assumed
that precipitator capacity will have to be in-
creased by 120% (see pp. 30-31).
Waste Disposal
In estimating the cost of additional
space for waste solids disposal) it is as-
sumed that enough area will be procured to
store the solids for 10 years of plant opera-
tion. This is a typical practice in the wet-
process phosphoric acid industry) where waste
byproduct calcium sulfate also constitutes a
major disposal problem.
Stream Pollution
Since calcium sulfate is relatively in-
soluble) it is assumed that sluice water carry-
ing the waste solids to disposal can be drain-
ed directly to running streams. This is gen-
eral practice in the phosphoric acid industry.l
The main criterion seems to be the degree of
dilution afforded by the stream or other body
1 Such water in the phosphate industry is
treated with lime) but this is for neutrali-
zation of acidity and has nothing to do with
the calcium sulfate.
of water into which the sluice water is emit-
ted. Most of the central power stations) be-
cause of the need for large amounts of cooling
water) appear to be located adjacent to bodies
of water large enough to give the necessary di.
lution. Where this is not feasible) it will
be necessary to recycle the pond water as is
done in some large phosphoric acid plants (to
avoid the cost of water treatment to neutral-
ize acid). Since this probably would only be
necessary for a few plants) the cost of recy-
cling has not been estimated.
If dolomite is used) the soluble magne-
sium sulfate formed would aggravate the pollu-
tion problem and increase the number of plants
for which sluice water recycling would be nec-
essary.
The main stream pollution problem may be
from the unreacted limestone rather than from
the calcium or magnesium sulfate. This will
enter the pond as calcium hydroxide and will
raise the pH of the effluent to a high level.
There is not enough information available to
evaluate the problem; for locations where dis-
posal to streams is infeasible) the sluice
water will have to be recycled or dry transfer
to the disposal area adopted.
Control Function
It has been considered that limestone in-
jection may be used by power plants in two
ways--continuously where local conditions dic-
tate and occasionally where the main need is
to avoid high ground concentrations that de-
velop only when atmospheric conditions are ad-
verse to dispersion from the stacks. The
costs of both types of operation have been
estimated. For incident control) it is as-
sumed that the injection system would be oper-
ated ten times annually for periods of 3 days
each.
23
-------
PROCESS EQUIPMENT
Equipment needs in limestone injection
are mainly related to receiving) storing) and
grinding tbe incoming limestone. Altbougb in-
jection is tbe important process step) equip-
ment needs for it are minor.
There are several alternative cboices in
selection of tbe limestone bandling and grind-
ing equipment.
Major Alternatives
Particle Size of Purcbased Sorbent: The
first cboice affecting equipment selection is
tbe particle size of sorbent purcbased. If
bydrated lime is purcbased) tbe small particle
size would dictate use of special trucks) rail
cars or barges) and pneumatic bandling systems
at tbe power plant. Raw or calcined limestone
could be purchased eitber ground or in lump
form. Because of tbe care) and perbaps some
additional expense) required for sbipping fine-
ly ground material) purcbase of lump material
seems best. For the large amounts involved)
tbe grinding can be done in most instances as
cbeaply at tbe power plant as at tbe limestone
supplier's plant) witb tbe advantages of sim-
plifying sbipment and power plant storage.
Storage at Power Plant: Calcined lime-
stone would bave to be stored in closed equip-
ment) of course) but raw limestone offers a
choice. The incoming material could be stored
either in a storage silo or in a combination
of silo and pile storage. The silo is essen-
tial in any event; even if pile storage were
selected) the silo would be needed to serve
the same purpose as the bunkers in the coal-
feeding operation) that of giving some leeway
in scheduling the workers wbo reclaim from the
storage pile and move coal to the bunkers.
For a single medium-size boiler such as
the 200-megawatt unit in the present study) it
seems unlikely that the expense of a system
similar to coal handling--delivery of incoming
material to a pile) reclaiming from the pile)
and transfer to bunkers--need be incurred. De-
livery direct to the silo would be simpler and
less expensive. A silo bolding 2 or 3 days'
supply would allow operation over a weekend
without supply and should provide an adequate
reserve for incident control over a period of
high air pollution potential. It would be es-
sential to have a quickly available supply on
hand to ensure against an interruption in de-
li very.
To guard against extended interruptions
in delivery) a "dead" storage pile could be
maintained. Since this would be used only sel-
dom) and would be reclaimed on an emergency
basis with the plant coal-handling facilities)
24
for capital cost and operation of
equipment has been included in the
no expense
reclaiming
estimates.
An incidental advantage of delivery direct
to a silo is that the limestone would be kept
dry. In many instances) limestone could be
purchased with a guaranteed moisture content
of only 1 to 2~. Keeping such material dry
during storage would reduce drying costs and
avoid trouble from wetting and freezing in the
winter.
Even for a 200-megawatt unit) a silo large
enough for a 2-day supply would hold from 200
to 800 tons (range for 2-3.5~ S in coal and 1-2
times stoichiometric injection). Thus tbe size
of delivery units by truck or rail (about 20
tons/truck load and 50 tons/rail car) would not
be a problem. For barge or unit train delivery)
however) silos large enough to hold a full ship-
ment might be impractical and a system similar
to coal handling would be preferable.
Grinding System: Selection of a grinding
system depends mainly on the particle size re-
quired and the amount of limestone to be ground.
For the combination of a 200-megawatt boiler
and a particle size of 70% through 200 mesh
(the larger size assumed for the estimates)) a
ring-roll or ball-race mill of tbe type normal-
ly used in power plants to grind coal should be
quite suitable. Such a mill has been assumed
for the detailed cost estimates.
If grinding to very fine particle size is
found to be more economical than using more
limestone) a ring-roll mill should be suitable.
Such mills are used commercially to grind lime-
stone to a particle size as low as 99.5% minus
325 mesh (44 microns). Ball mills would also
be acceptable; one company grinds limestone in
a ball mill to a particle size of 90% minus 40
microns (about 350 mesh)) 55~ minus 10 microns)
and 10~ minus 2.5 microns.
If even finer particle sizes were consid-
ered desirable) fluid energy mills might be
considered. It has been reported that such
mills are suitable for grinding limestone to
an average particle size of 3.5 microns. In
the course of the present survey) however) it
was learned from one of the leading fluid
energy mill manufacturers that this type of
grinding equipment is not applicable to lime-
stone. The physical properties of the lime-
stone are not suitable.
The amount of limestone to be ground
appears to be a more important factor than
particle size in mill-type selection. The
grinding requirement for the 200-megawatt unit
considered in this study would range between
5 and 18 tons per hour) well within the range
-------
of a single ring-roller mill at the coarser
particle size. For larger units) however) use
of this type would require a multiple-mill sys-
tem as ring-roll mill capacity seldom exceeds
30 tons per hour (at a 70% -200 mesh grind).
For example) the limestone required for a 1500-
megawatt plant) assuming 3.5% sulfur in the
coal and twice stoichiometric injection) would
be over 130 tons per hour. For such a grind-
ing rate) a ball milll would have lower first
cost than the several roller mills that would
be required. Operating costs may also be lower;
in the grinding of phosphate rock) a material
comparable with limestone in grindability) it
has been found that ball mills give the lowest
cost for rates over about 50 tons per hour.
If extremely fine grinding were required)
the ball mill type would take on an additional
advantage because the reduced output resulting
from the finer grinding would make a multiple-
roller mill installation necessary at even
lower limestone consumption.
Extremely fine grinding may require con-
sideration of methods such as use of grinding
additives and of wet grinding. The latter) for
example) has been reported to reduce power re-
quirement and increase capacity considerably)
as compared with dry grinding. Wet grinding
normally is more expensive overall because of
the cost of drying. For limestone injection)
however) it may be feasible to inject the slur-
ry directly from the mill. As noted earlier)
slurry injection is being studied in Germany,
primarily because of ease of handling.
The method of classification is also a
consideration in designing the milling system.
A closed circuit is required) but various clas-
sifier arrangements can be used. In coal and
phosphate rock grinding) the mill usually has
an internal classifier with an air sweep to re-
move particles as they become reduced to the
desired size. Alternately) the classifier can
be located externally to the mill and the par-
ticles conveyed to it by the sweeping air. Or
the mill can be discharged by gravity and the
material transferred to the classifier by a
conveyor such as a bucket elevator.
The air-swept) internal classifier type of
mill works well for phosphate rock a~d coal in
grinding to a range such as 70% minus 200 mesh.
Also) the ring-roller version of this type is
reported by one of the mill manufacturers to be
suitable for 325-mesh grinding. Therefore) in-
ternal classification has been assumed for the
present study. For larger installations such
as the 130 tons per hour mentioned earlier) the
1 The 1500-megawatt plant is assumed to be a
multiple-boiler plant) for which a central
grinding unit with a single mill would be
more economical than a mill for each boiler.
use of ball mills would make external classi-
fication preferable. Gravity discharge and
bucket elevators in such a system would recuce
power consumption as compared with conveying
the ground material to the classifier with the
air sweep; offsetting this is the difficulty
of avoiding a dust problem in operating and
maintaining the elevators. Moreover) power is
relatively inexpensive at a power plant.
Choice between air sweep and gravity dis-
charge might well depend on local conditions
at the power plant.
Limestone Drying: The main alternative
in limestone drying is between using heated
air from the air preheater) which is normal
practice in drying coal) and installing a sep-
arate fired heater to heat the drying air
(either mill-sweeping air or classifier air if
the mill is not air swept). For the base study
of the 200-megawatt unit) the usual practice
has been assumed. However) if a central grind-
ing unit were installed for a multiple-boiler
installation) it would probably be better to
use a fired heater.
Ground Limestone Storage: A storage silo
for ground limestone would increase operating
flexibility but is probably not justified for
a 200-megawatt unit. The surge hopper just
before the mill should give adequate flexibil-
ity and allow feeding the limestone from the
mill to the boiler without major storage be-
tween.
For a large installation) however) feed-
ing directly to the boilers from a mill such
as the 130-ton-per-hour unit mentioned earlier
probably would not be practical and a ground
rock silo would be required.
Metering and Injection: Coal is usually
metered in the unground state) just before the
mill) and the ground material carried directly
into the boiler with the sweeping air. In
such a system) the metering must be done on
the coarse material for the fines are never
separated from the gas stream after the mill.
This system is undesirable for limestone in-
jection because of the relatively large amount
of sweeping air (about 2.0-2.5 lb. air/lb. of
solids) that would be injected into the boiler.
Therefore) even if air is used to sweep the
mill it must be separated from the ground lime-
stone) for example) by a cyclone and bag fil-
ter) and the fines then reinjected into a mini-
mum amount of conveying air. Hence the fines
are available for metering if desired. This
is of some advantage) as it allows volumetric
metering) which) although not highly accurate)
is a simple and inexpensive method. It can be
used because limestone injection does not re-
quire a high degree of accuracy.
The metering and injection system selected
for the 200-megawatt boiler) then) involves
25
-------
grinding in an air-swept mill) separating and
collecting the fines in a feed bin fitted with
a bag filter) and metering from the bin by a
screw feeder delivering into a rotary valve
that delivers the material into an air-injec-
tion tube.
This system is more complicated than the
relatively simple coal-feeding system. It
might be feasible to use the latter for lime
injection if recirculated stack gas could be
used rather than air) to avoid the heat loss
that the air produces by increasing the volume
of gas leaving the system. The main diffi-
culty is adverse effect on the thermal balance
and pressure drop in the boiler. Even if gas
were taken just before the air heater (to avoid
any effect on heat recovery in the air heater))
cleaned) and recirculated to the mill as sweep-
ing gas) the large amount required might upset
boiler operation unless the boiler were design-
ed for it. Hence the method is questionable
for existing plants but might be advantageous-
ly designed into new plants.
Injection with Coal: If it should be
found that injection with the coal is desirable)
a choice among feeding systems would have to be
made. If a particle size no smaller than that
of the coal were acceptable) the best course
probably would be to meter coarse limestone
along with coal to a mill and grind them both
together. The sweeping air would then carry
the limestone into the boiler along with the
coal.
If a finer size were required) the lime-
stone would have to be ground in a separate
mill. A simple means of injection would be to
meter this into the coal-air mixture just be-
fore it entered the boiler. In either case)
the objective of using the momentum of the air
and coal entering the boiler to carry the lime-
stone in and distribute it uniformly over the
boiler would be realized.
Although feeding with the coal should help
in sorbent distribution) little reduction in
investment or operating costs would be obtain-
ed. Therefore) separate estimates for injec-
tion with coal have not been included.
Equipment Description
For the plant arrangement selected for the
200-megawatt unit (Appendix V)) limestone is
received by truck shipment) dumped into an un-
loading hopper) and conveyed to the storage
silo. From storage) the material is conveyed
to a surge hopper over the mill) ground) blown
by the mill-sweeping air to a feed bin equipped
with a bag-filter vent) passed through a vari-
able-speed screw feeder) and fed into the air-
injection pipe by a rotary valve. The bin is
mounted on a load cell to allow calibration of
the feeder.
26
Other capital cost items needed are addi-
tional precipitator sections and ash-storage
pond area.
The base case assumed for the equipment
design is 3.5% sulfur in the coal and twice
stoichiometric injection. This combination
was selected because the limestone consumption
would be the highest of the four combinations
of 2.0 and 3.5% sulfur coal with 100 and 200%
stoichiometric injection (proportions are 7) 4)
3.5) and 2)) and the equipment therefore would
give the most problems in fitting into a power
plant because it would be the maximum size.
Smaller equipment for the other combinations
should not offer any problemsj investment for
smaller systems can be obtained from Figure 12.
Unloading Hopper: The unloading hopper
is a 20- by 20-foot concrete hopper with sides
sloping to a center opening above a belt con-
veyor 12 feet below gradej the top of the hop-
per is at grade for dump truck receipts. The
hopper will accommodate about 50 tons of lime-
stone or about three truckloads.
Unloading Conveyor: The unloading con-
veyor transports stone from the unloading hop-
per to a storage silo. The conveyor is a
truss-frame) troughed-belt type) 180 feet in
length with a 24-inch belt.
Storage Silo: Incoming limestone is stor-
ed in a silo made of mild steel. The unit is
35 feet in diameter with 40-foot straight sides
and a 16-foot conical bottom. The silo will
accommodate 1000 tons of limestone or about a
60-hour supply for 200% stoichiometric treat-
ment with 3.5% sulfur in the coal (17.5 tons/
hr.)j this reserve is sufficient for operation
over a weekend without delivery. The storage
silo discharges onto a transfer conveyor belt.
Transfer Belt: Limestone is conveyed from
the storage silo to a surge hopper by a 230-
foot belt of identical construction to the un-
loading conveyor.
Surge Hopper: A surge hopper is provided
above the mill to protect the injection system
from minor delays in operation of the transfer
system. The hopper is a mild steel tank 18
feet in diameter with 20-foot straight sides
and a 12-foot conical bottom. The storage ca-
pacity is equivalent to 8 hours' supply.
Limestone Feeder: Rate of feeding lime-
stone to the mill is controlled with a belt-
type feeder with a capacity of 20 tons per hour
Pulverizer: The pulverizer provided for
reducing the size of the limestone from minus
1/4 inch to 70% minus 200 mesh is a Band W
type E bowl mill of the type used for grinding
coal. The mill has a capacity of 20 tons per
hour when grinding coal to 70% minus 200 mesh
(adequate for twice stoichiometric injection
with 3.5i S coal). Air supply for classi-
fication and drying in the mill and transport-
-------
ing pulverized limestone to the feed hopper is
pulled from the preheated air duct which sup-
plies air for combustion.
Ground Limestone Feed Tank: The ground
limestone feed tank serves both as a receiver
for the mill and as a supply for the injection
system. The tank is a mild steel cone-bottom-
ed tank 10 feet in diameter and 15 feet high
equipped with a bin vent-type bag filter to
handle 500 cubic feet per minute of air. An
exhaust fan is provided to compensate for the
pressure drop across the filter. The tank is
mounted on load cells for rate checks.
Injection System: Ground limestone is
metered into the injection system through a 9-
inch screw feeder equipped with a variable-
speed drive; capacity of the injection system
is 20 tons per hour. The feeder discharges
through an 8-inch rotary valve into a mixing
nozzle. This single metering system supplies
sorbent for coal delivered from six coal
scales. Instrumentation for ratio control is
provided. The limestone is transported in
dense phase (0.07-0.08 lb. air/lb. of solids)
through a flow splitter to the injection point.
The flow splitter is a cylindrical tank with
six 2-1/2-inch pipe outlets equipped with ori-
fice plates. The flow of air and entrained
solids is divided into six streams and convey-
ed through separate 2-1/2-inch pipes to flow
nozzles that project through the boiler wall;
the nozzles can be pivoted in a vertical plane
to vary the angle of injection. The velocity
and angle of injection will be determined ex-
perimentally. Separate rotary compressors
(maximum of 10 p.s.i.g. discharge pressure)
will provide up to 600 cubic feet per minute
each for conveying and for injection of the
limestone; the injection air will be added to
the transport stream in mixing tubes just ahead
of each nozzle.
Precipitators: To prevent an increase in
dust emission from the unit as a result of
higher dust loading, additional precipitator
capacity is necessary. The normal dust load-
ing of 0.7 grain per cubic foot (after the me-
chanical collector) with no limestone injec-
tion would increase to 1.8 grains per cubic
foot with 200% stoichiometric addition of lime-
stone for 3.5% sulfur coal. To maintain an
emission level of 0.07 grain per cubic foot,
it was necessary to add two 18-foot precipita-
tor sections in series with the original 20
feet of treatment sections. With addition in
series, the velocity remained at 6.5 feet per
second but the residence time in the treatment
sections increased from 2.3 seconds to 6.4
seconds.
Dust Handling: The dust-handling system
from the precipitator hoppers to the storage
pond was modified to accommodate the additional
load imposed by limestone injection. The ex-
isting system includes pneumatic conveyors to
transfer dust from the precipitator hoppers to
a central ash sluice point from which sluice
pipes carry the dust as a 12% solids water
slurry to a settling pond about 1/4 mile away.
Additional equipment required consists of:
1. A 1000-gallon-per-minute sluice water
supply pump with 1500 feet of 8-inch piping.
2. A parallel pneumatic conveyor with an
additional compressor.
3. A parallel 12-inch sluice pipe to the
existing storage pond.
Storage Pond: The impounded volume for
disposal of ash was increased to accommodate
the additional solids load resulting from lime-
stone injection. It was assumed that real es-
tate would be available at the price paid for
the steam plant property. During continuous
operation with injection of 200% stoichiometric
limestone for 3.5% sulfur coal, 60 acre-feet
will be required annually for the additional
dust. Assuming impoundment to a height of 10
feet, which is current practice for fly ash at
Colbert, 6 acres per year additional space is
needed. Cost of pond development for 10 years
of operation is included in the investment es-
timate.
27
-------
EFFECT OF LIMESTONE INJECTION ON POWER PLANT OPERATION
Results of limestone injection tests in
pilot and small-scale plant equipment, together
with a general knowledge of power plant opera-
tion, provide a basis for evaluating possible
operating problems which might develop as a re-
sult of sulfur dioxide control by dry injection.
Further testing of the process in large-scale
equipment is necessary to quantify both the ad-
vantages and disadvantages during routine oper-
ation of a base-loaded unit. Possible effects
of limestone injection on performance of a
power boiler and auxiliary facilities are dis-
cussed in the following sections.
Combustion
Conversion of chemical energy to electri-
cal energy in a high-performance boiler and
generator by combustion of coal has been de-
veloped into a highly reliable operation.
Methods for grinding coal for desired particle
retention time, controlling air to coal ratio
for efficient combustion, and dispersion of
combustion products for optimum heat transfer
and steam temperature control have been adapted
to various boiler types. Introduction of a
foreign material such as limestone will re-
quire some adjustments in firing techniques to
compensate for changes in heat effects.
Regardless of method of injection, the
heat required to raise the temperature of the
additive and the heat necessary to dissociate
the carbonates must be supplied from combus-
tion of coal. A discussion of the thermal
effects on operating costs is presented in
Appendix IV. In some installations, the gen-
erating capacity limits the boiler capacity;
in these plants additional fuel can be burned
to supply the added heat requirement. In many
plants, however, the boiler is fully loaded
and injection of limestone will require sacri-
fice of capacity equivalent to the added heat
requirement.
The method of injection will influence
the effect of injection on combustion. Addi-
tion of limestone admixed with coal could
cause flame stability problems. Heat transfer
to the sorbent within the same boundary nor-
mally occupied by the burning coal-air mixture
will reduce to some extent the driving force
for combustion, and the carbon dioxide liber-
ated1 during calcination will dilute the oxygen
needed to maintain a high-intensity flame.
During plant tests at Colbert Steam Plant, a
10% limestone-90% coal mixture (stoichiometric
for 4% S in coal) was burned without flame sta-
bility problems; tests at Wisconsin Electric
1 About 4% more carbon dioxide than usual.
28
resulted in the same observation. It was the
opinion of the TVA power production personnel
who assisted with the tests that doubling the
dosage would have caused combustion problems.
To assure flame continuity, it will probably
be necessary to omit injection during boiler
light-off and shutdown.
Effects on combustion of addition of
limestone with the secondary air would be sim-
ilar to those discussed for coal-limestone mix-
ture. A slight difference might be better
availability of oxygen because of better dis-
persion of limestone with respect to coal.
Injection of limestone above the burners
would likely have little effect on combustion.
However, if limestone is injected with a car-
rier fluid such as air or recirculated gas, the
gas flow pattern and therefore the heat trans-
fer may be changed. Also, additional fuel
equivalent to the heat required to raise the
temperature of the carrier fluid will be needed.
Slagging and Heat Transfer
In addition to possible effects on combus-
tion, limestone injection may influence heat
transfer by changing the amount, texture, and
location of deposits on tube surfaces. Coal-
ash corrosion that occurs on tube surfaces ex-
posed to high temperature is a result of metal
attack by complex salts containing sulfur ox-
ides.2 Reaction of limestone with the sulfur
oxides should reduce the corrosive attack.3
On the other hand, reaction of calcium
with the fly ash may result in altering the
slag composition and change ash-softening tem-
peratures. A difference in softening point
would change the location and extent of slag
deposits on tube surfaces.
Results of tests at Detroit Edison (Appen-
dix I) showed that ash-fusion temperatures
dropped about 2500 F. for ash-dolomite mixtures
in which dolomite represented 10 to 30% of the
mixture. Above 30% dolomite, the ash-fusion
temperature increased again and equaled the
"ash only" value at about 46% dolomite; at this
dolomite level, the ash contained 24.4% calci-
um oxide and 17.7% magnesium oxide. The re-
£ Jonakin, J., Rice, G. A., and Reese, J. T.
"Fireside Corrosion of Superheater and Re-
heater Tubing," ASME Paper 59-FU-5.
3 Borio, R. W., Ulmer, R. C., Hensel, R. P.,
and Wilson, E. B. "Study of Means for Elim-
inating Corrosiveness of Coal to High-Temper-
ature Surfaces of Steam Generating Units, If
ASME Paper 67 WA-CD-3.
-------
suIts of earlier work by Nicholls and Reidl
showed a similar reduction and subsequent in-
crease in ash-softening temperature with in-
creasing calcium content (Fig. 5).
2700
\
2600
\
Ii:
.
I&i' 2500
a:
::)
~
~ 2400
Q.
::e
IIJ
I-
~ 2300
o
..J
IL
I
I
I
I
2200
2100
o
40
50
10
20 30
CoO IN ASH, %
Figure 5. Effect of Lime Content on
Fusibility of Coal Ash
{Nicholls) P.) and Reid) W. T. 'Irans.
ASME 54, 167-190 (193217 -----
Plant tests at Wisconsin Electric result-
ed in unusual slag accumulation. The deposits
were quite tenacious at operating temperatures
and were not easily removed with conventional
soot-blowing equipment. However) when cool
the material was quite friablp. and easily bro-
ken. During the plant tests by TVA,deposits
were considered normal except in texture and
were removed in the conventional manner. Zent-
graf noted that deposits formed during lime-
stone injection were friable and easily remov-
ed by soot blowers although more frequent
cleaning was required. Prolonged exposure of
slag accumulation to high temperature formed
a hard sinter. He also found that lead tubes
in the superheater section were scrubbed clean
by impingement of fly ash.
It is important that the ash-softening
temperature be above the gas temperature enter-
ing the convection pass so as to avoid any un-
due accumulation of slag on the closely spaced
tubes in the convection zone. The gas temper-
l Nicholls) P.) and Reid) W. T.
54) 167-190 (1932).
Trans. ASME)
ature at this point will depend on furnace
heat sorption) which in turn will depend on
ash deposits on the furnace walls. The ten-
dency toward adverse fouling will therefore
be related to the ash-softening temperature
characteristics of the coal ash and limestone
mixture and furnace heat sorption efficiency.
Some control of slagging may be effected by
adjustment of total air and burner regulation.
Once these are optimized) it is then necessary
to depend on furnace and convection pass
blowers to remove excessive deposits. After
slagging patterns have been established) it
may be necessary to modify) relocate) or add
blowers for this purpose.
Detrimental effects of slagging may in-
clude difficulty in maintaining steam temper-
ature control) an increase in exit gas temper-
ature) reduced boiler availability) reduced
peaking capability) an increase in use of
steam or air for boiler cleaning) and an in-
crease in gas fan power. Most of these effects
would reduce boiler efficiency and increase
operating costs.
Erosion
Loss of metal from boiler heat-transfer
surfaces as a result of impingement of fly ash
particles is a common problem. Since lime-
stone injection will increase the concentration
of dust particles in the gas stream) acceler-
ated wear can be expected. Limestone has a
hardness index of about 3 on Moh's scale; coal
has an index of 0.5 to 2.5 and fly ash is some-
what harder. The particle-size range for both
the injected limestone and the calcium sulfate
will be comparable with the fly ash) so the mo-
mentum of particles will also be similar.
Abrasion protection may be required.
Tube shielding to reduce erosive effect
of fly ash is common practice. A high-temper-
ature stainless steel such as Type 309 is
needed to resist temperatures at superheater
and reheater tube locations. In areas where
soot blowers are required) shielding is often
used to protect surfaces from high-velocity
particles entrained in the steam or air emit-
ted from the blower. If more frequent blowing
is required to control deposits during lime-
stone injection) shielding costs will probably
increase. High-temperature shielding costs
about $2.25 per lineal foot.
Shields are usuall~ installed only after
visual inspection and tube measurements show
evidence of metal loss. They are normally in-
stalled to protect front rows of tubes and
along paths of heavy ash burden. Operating
experience will be needed to assign shielding
costs attributable to limestone injection.
29
-------
Air Heater Operation
Most boilers burning high-sulfur coal ex-
perience low-temperature corrosion and plugging
of air heaters. The sulfur trioxide produced
during combustion reacts with water vapor to
form sulfuric acid which condenses on air heat-
er surfaces. The moist acid film promotes de-
posit buildup and eventual plugging of air
heater passages.
Air heaters are washed periodically to
remove deposits. The sulfate salts present in
the deposits hydrolyze to form sulfuric acid.
The sulfuric acid in turn produces a solution
which is very corrosive to air heater surfaces.
Combustion Engineering, in its work at
Detroit Edison, reports that the addition of
limestone completely neutralizes sulfur tri-
oxide in flue gas. This has also been the ex-
perience of TVA in short-term tests. It is
reasonable to expect that air heater corrosion
and plugging will be significantly reduced if
sulfur trioxide is eliminated from flue gas.
Thus, maintenance savings in the form of re-
duced air heater corrosion and plugging can be
credited to limestone addition. Combustion
Engineering allowed a credit of $0.053 per ton
of coal in cost estimates of corrosion savings.
In the estimate made in conjunction with this
report, the corrosion credit assumed was only
about 10% of this amount.
The removal of sulfur trioxide from flue
gas might also permit a reduction of flue gas
temperature leaving the air heater with result-
ing savings in cycle efficiency.
Dust Collection
Based on a 200-megawatt pulverized-fuel-
fired unit burning 3.5% sulfur fuel with 12%
ash, calculated ash burdens are shown in Table
IV for the 100 and 200% stoichiometric injec-
tion rate (assuming 30 and 50% S02 removal,
respectively). In view of the increase in ash
burdeD} consideration must be given to the ef-
fect that increased dust loading will have on
dust collection, handling, and disposal. The
major equipment items affected are mechanical
and electrostatic fly ash collectors and the
dust-handling system.
Table IV. Ash Burden for a 200-Megawatt
Pulverized-Fuel-Fired Unit, Tons Per Hour
Without
limestone
Stoichiometric ratio
1.0 2.0
Bottom ash
Fly ash
4.0
12.0
5.5
16.5
2.31
6.94
Tbtal
16.0
9.25
22.0
30
Mechanical Collectors: A sizable per-
centage of older boilers are equipped with me-
chanical collectors. The efficiency of a
mechanical collector is fairly constant over
a wide range of grain loading, assuming par-
ticle sizes above 5 microns. Collection effi-
ciency is poor on particle sizes below 5 mi-
crons.
Since limestone and fly ash particles are
expected to be larger than this, the increase
in grain loading due to limestone addition will
not affect efficiency. However, the outlet
grain loading will increase in proportion to
increase in inlet loading.
An increase in ash burden to a mechanical
collector will contribute significantly to
collector maintenance costs. Impingement wear
on the inlet and erosion of the collecting
tubes and vanes account for the major portion
of the maintenance. Typical maintenance costs
for mechanical collectors on a 200-megawatt
pulverized-fuel-fired unit burning 12% ash
coal range from $0.005 to $0.029 per ton of
coal burned. Maintenance costs are expected
to increase in direct proportion to the in-
crease in ash burden.
Electrostatic Precipitators: Test work
conducted at Detroit Edison's St. Clair plantl
indicates that injection of dolomite seriously
affects the performance of electrostatic pre-
cipitators. The addition of 1.1 times the
stoichiometric quantity of dolomite reduced
precipitator efficiency from 80 to 55%. The
ash burden going to the precipitator was about
doubled. The net result of the increased
loading and decreased efficiency was a four-
fold increase in dust emission. Reduced effi-
ciency was attributed to change in dust resis-
tivity, one of the main criteria for precipi-
tator design. Most of the sulfur in coal is
oxidized to sulfur dioxide during combustion
but a small amount (about 0.5%) forms sulfur
trioxide. The sulfur trioxide content in-
fluences resistivity of the dust, and is taken
into account in precipitator design. Removal
of sulfur trioxide from the gas stream by sorp-
tion on limestone probably increases resistiv-
ity beyond the level for efficient operation
of a precipitator designed for use on a unit
burning high-sulfur coal. A plot of precipi-
tation rate as a function of fly ash resistiv-
ity is shown in Figure 6. The effect of resis-
tivity is particularly significant at the oper-
ating temperatures (about 3000 F.) normally
encountered in precipitators installed on power
boilers. At higher temperatures, variation in
dust resistivity would have less effect on pre-
cipitator efficiency; if heat recuperation
l See Appendix I for full discussion.
-------
downstream of a precipitator could be done
practically, efficient precipitation probably
could be carried out with little concern about
sulfur trioxide level in the combustion gas.
~ 0.6
o
u
1&1
U)
a:: 0.!5
1&1
0.
I-
~ 0.4
...
---
-....
~
1&1 0.3
...
-------
storage of the solids added by 200% stoichio-
metric treatment of 3.5% sulfur coal.
Water Pollution
In a typical thermal power plant, water
used to sluice fly ash to storage is discharged
to a water course after the solids have settled
from it. Because of the large requirement of
water for condenser operation, power plants are
normally located on sizable streams. The ef-
fluent from ash ponds is usually discharged
over a skimmer device to prevent discharge of
floating ash or other foreign debris. Disper-
sion of the discharge stream generally is not
a problem and dissolved solids are diluted to
an acceptable level.
Injection of limestone into a power plant
boiler will contribute to water quality control
problems in two ways. First, because of the
higher dust burden, additional sluice water
will be required. Secondly, solubility 01' re-
action products may result in higher concen-
trations of dissolved solids. The suspended
solids concentration when sluicing lime-dust
mixtures is assumed to be the same as for fly
ash, 12%.
For 200% stoichiometric injection of lime-
stone and 3.5% sulfur coal with 12% ash, the
volume of water required for sluicing the dust
yill be about three times that for normal fly
ash or, theoretically, 630 gallons per ton of
coal versus 220 gallons. Distribution of this
added volume of effluent in the receiving water
course will require closer surveillance.
Several water quality parameters may be
affected by the limestone injection process,
including color, pH, hardness, suspended solids,
and dissolved solids. Because of a more gen-
eral awareness of the necessity for water qual-
ity control, regulatory agencies are becoming
more stringent on permissible concentrations
in receiving streams. For example, the recom-
mended limit for sulfate in drinking water is
250 p.p.m.~ but recent proposals have included
reduction of sulfate to 150 p.p.m. Present
limit on total dissolved solids is 500 p.p.m.
The sluice water from storage ponds con-
taining limestone products from an injection
system will have a relatively high concentra-
tion of dissolved solids. Depending on lime-
stone composition, degree of reaction of lime-
stone with sulfur dioxide, extent of reaction
of contained magnesium oxide, required sluice
water slurry concentration, and temperature of
pond water, the effluent might contain as much
as 6000 p.p.m. S04-' 1600 p.p.m. ea+, and 1100
p.p.m. Mg+ contributed from limestone reaction
~ Public Health Service Publication
U. S. Government Printing Office,
Washington, D. C. (1962).
No. 956,
32
products. A high degree of dilution, up to 50
to 1, will be necessary to avoid excessive con-
centration in the receiving stream.
The alkalinity of the pond effluent could
be a problem. The flow from ponds containing
only fly ash is alkaline; the pH will be fur-
ther increased by unreacted lime from the in-
jection process. If all of the added calcium
oxide reacted with sulfur dioxide in the boil-
er; the pH of the sluice water would not be
significantly changed. However, the sorbent
will not be fully utilized and any unreacted
calcium oxide will form calcium hydroxide when
contacted by sluice water.
Information from plant-scale studies is
needed before the impact of limestone injection
on water quality control practice can be pre-
dicted accurately. The conditions will vary
widely from plant to plant depending on sulfur
and ash content of coal, type of limestone
available, effluent distribution, and size of
receiving stream.
Stack Height
Since limestone inu=ction with dry collec-
tion will not significantly alter either the
temperature or volume of gas emitted, it should
not significantly alter the plume trajectory.
It, therefore, has no direct bearing on stack
design. Several factors, however, suggest that
stack design and heights might be altered as a
result of limestone injection.
High stacks normally result in effective
dispersion of the gas and thus reduce ground-
level concentrations of sulfur dioxide. With
removal of sulfur dioxide from the gas stream,
such dispersion would not be necessary to main-
tain a given ground-level concentration. One
might thus surmise that a shorter stack would
suffice. The stacks, however, must also dis-
perse the other constituents of the stack gas,
such as oxides of nitrogen and uncollected fly
ash, to prevent their ground-level concentra-
tions from becoming objectionable. The na-
tion's current awareness of air pollution prob-
lems is likely to lead to continuous lowering
of admissible ground levels of sulfur dioxide
and other stack gas constituents, and thus may
require both sulfur dioxide removal and tall
stacks. One cannot be optimistic that a de-
crease in capital costs attributable to short-
er stacks might be realized from limestone in-
jection.
Removal of sulfur dioxide results in de-
creasing the corrosive nature of stack gases.
This may lead to consideration of gains in
thermal efficiency of lowering exit gas temper-
atures. A lowering of the temperature reduces
both the buoyancy and momentum components of
plume rise. The efficacy of these components
in altering the rise of the plume above the
-------
stack is currently under intensive study. A
lowering of exit gas temperature might result
in consideration of changing exit velocity or
increasing stack height. An economic balance
between gains from thermal efficiency and in-
crease in capital cost resulting from change
in stack design may be required to determine
the amount the exit temperature could be re-
duced.
The trend for large central stations is
to tall stacks and several now exist which
equal or exceed 800 feet. Stacks to be com-
pleted in 1969 for coal-burning power plants
will average 609 feet (1960 average was 243
ft.) according to a recent survey. 1 It is not
likely that limestone injection can reverse
this trend.
1 "Proceedings: The
on Air Pollution,"
p. 152.
Third National Conference
December 12-14, 1966,
33
-------
ECONOMIC EVALUATION
The major cost factors involved in lime-
stone injection are summarized in this section
and curves relating the factors are presented.
Detailed cost estimates are given in Appendix
IV.
Sorbent Cost
Vendor's Price: The f.o.b. cost of the
limestone supply assumed, from a quarry near
the Colbert power plant, has been quoted by
the vendor at $1.35 per ton. Trucking cost to
the power plant would be $0.70, giving a de-
livered cost of $2.05 per ton.
In addition, information on limestone cost
in other areas and of calcined limestone and hy-
drated lime, was obtained by a survey of lime
manufacturers, located mainly in the eastern
part of the country. The prices quoted are
given in Table V.
It should be noted that these quoted
prices, as well as quoted figures in later sec-
tions, are based on costs in existing plants
and on an amount of material (50,000 tons/yr.)
considerably less than that which would be used
by a modern,large power plant. If a long-term
contract for a very large quantity were entered
into, then the vendor might be justified in
putting in new equipment and quoting a special
low price.
It is not possible, of course, to estimate
very closely the purchase price under a long-
term contract for a very large quantity of
limestone, say, the 700,000 tons per year need-
ed for a 1000-megawatt unit (at 3.5% S in the
coal and twice stoichiometric). A good figure
could be obtained only by actual negotiations
in a specific situation. However, discussions
with limestone producers have supplied a rough
estimate of the potential decrease. The con-
sensus for a 700,000-ton-per-year supply was
about $1.15 per ton; the range was $0.71 to
$1.50 per ton.
It is concluded that for a relatively
small limestone consumption, for example, the
40,000 tons per year needed for a 200-megawatt
unit with 2% sulfur coal and stoichiometric in-
jection, or the 33,000 tons needed for a 1000-
megawatt unit under the same conditions but
operated on an intermittent injection basis,
the quoted price range of crushed limestone
(maximum size, 1/4 in. or larger), $1.35 to
$2.00, is a good figure. For larger cons~~p-
tion, however, particularly for the 1000- to
2000-megawatt or larger central stations, much
lower cost can be expected. The ultimate might
be a situation in which a limestone plant of
very large capacity, located adjacent to a nav-
igable river, supplied several large power
34
plants up and downstream by barge; this is
similar to the supply situation on the Great
Lakes, where a large limestone company sells
crushed limestone in large quantities at $0.73
per ton f.o.b. ship.
Shipping Cost: Freight costs of sorbents
by rail or truck transportation are given in
Figure 7. Rail shipment is the most economi-
cal in most instances. Shipping costs in terms
of cost per ton of calcium oxide for the vari-
ous sorbents, by the lowest cost form of trans-
portation, are given in Figure 8.
Barging costs are much less than for rail
or truck but vary so widely that a curve of
cost versus distance would be misleading.
Barging is suited mainly to large-volume ship-
ment on a contracted minimum annual tonnage
basis, and, of course, to shipping and receiv-
ing points on or close to navigable water.
Where such a combination exists, very low rates
can be obtained; TVA, for example, ships coal
in large quantities at an average of about 2
mills per ton per mile. The conditions that
make this rate possible may be somewhat favor-
able, however; the following costs have been
estimated as generally applicable to barge
shipping of limestone:
Shipping cost/ton, $
Tons/yr. 100 mi. 200 mi. 300 mi.
150,000 0.50 0.75 0.90
1,000,000 0.45 0.65 0.80
There should be added to the barge rates
the cost of unloading at the power plant, since
there is little or no unloading cost for trucks
or hopper-bottom rail cars. A typical barge
unloading cost, based on power plant experi-
ence, is $0.16 to $0.17 per ton at an unload-
ing rate of 1000 tons per hour.
Shipping cost can also be reduced by the
unit-train method, that is, a special train
carrying only limestone and proceeding direct-
ly from quarry to power plant. This practice
is also generally based on a contracted mini-
mum annual tonnage and the cost varies widely
with conditions. TVA, for example, receives
coal by unit-train shipment at rates ranging
typically from 4.7 mills per ton per mile for
343-mile shipment to 9.2 mills for 114 miles.
Minimum annual tonnage for sUCh contracts is
1 to 2 million tons; minimum trainload tonnage
is about 5000 tons.
In comparison with Figure 7, which is
based on single carloads, the shipping cost for
the unit train is about 30% lower.
Grinding Cost: The particle size assumed
for the injected material in the subsequent
estimates is 70% minus 200 mesh (about 47%
-------
Table V.
Quoted Prices of Raw Limestone, Calcined Limestone, and Hydrated Lime
Location
Limestone
Cherokee, Ala.
Longview, Ala.
Ste. Genevieve, Mo.
Roberta, Ala.
Co1ema.n, Fla.
Russellville, KY.
Durbin, O.
K:lJnball ton, Va.
Stephens City, Va.
Strasburg, Va.
Adams, M F.o.b. cost/ton, bulk, $
indicated given mesh Cao JigQ Of material Of Cao and MgO
1/1+ in. 100 51.~ 1.7 1.35 2.55
200 70 53.5 0.8 3.85 7.10
16 100a 55.3 0.4 2.85 5.10
300 62b 55.3 0.4 2.85c 5.10
300 80 55.3 0.4 5.85 10.50
300 98 55.3 0.4 6.60 11.80
300 99.6 55.3 0.4 7.60 13.65
60 50d 54.0 1.65 3.05
325 75e 54.5 0.6 4.00 7.25
1/4 in. 100 51.7 3.6 2.00 3.60
3 in. 100f 30.6 21.5 1.50 2.90
200 48 30.6 21.5 2.50 4.80
200 82 54.0 2.0 3.00 5.35
200 75 54.8 0.3 2.75 5.00
3/8 x 1/8 in. 100 55.0 0.4 2.00 3.60
200 98g 54.0 0.7 10.00 18.30
200 99 55.0 0.3 7.00 12.60
5/8 x 1/4 in. 100 96.2 1.4 13.25 13.60
300 99 96.6 0.65 16.25 16.70
100 90 96.5 0.7 15.00 15.40
5/8 x 3/8 in. 100 96.5 0.7 13.50 13.90
3/8 in. 100 98.0 1.2 16.00 16.10
100 100 98.0 1.2 18.00 18.15
200 91h 57.6 40.0 12.50 12.80
- 92-97 14.00-18.00
10 50-7ji 99.0 0.1 15.00 15.15
1/8 in. 100 97.7 0.9 14.25 14.45
73.5 0.7 14.25 19.20
-k . 15.00
325 96.6 74.7 0.3 14.25 19.00
325 95 73.3 1.2 15.75 21.20
325 96.5 73.3 0.9 17.75 23.90
a 21% -80 mesh; sized product.
b 1001> -35 mesh.
c Low price due to product being surplus.
d 901> -10 mesh.
e 97.81> -100 meeh.
f 871> -40 mesh.
g 29% -15 microns; 1l~ -4 microns.
h 991> -100 mesh.
ji 80-90% -3 mesh; 0-10~ -100 mesh.
40% -100 .mesh.
k Not given. Presumed to be made fran Durbin,
0., calcined dolomitic l:1Jnestone above.
35
-------
10.00
RAIL - 100,000 LB. MINIMUM SHIPMENT
TRUCK-37,OOOLB. MINIMUM SHIPMENT
C
"'1.1.1
""'0..
(f)o..
0-
0:1:
en
..........
:1:«
C)-
i:ija::
a:: ~ 4.00
LI..<=;
~~
5:2:
m ~ 2.00
.....
~
8.00
TRUCK -ALL MATERIALS -
00
50
100 150 200
DISTANCE. MilES
300
Figure 7.
X
RAIL -CALCINED LIMESTONE
AND HYDRATED LI ME
RAIL - LIMESTONE
250
Bulk Transportation Cost for Shipping Limestone, Calcined
Limestone, and Hydrated Lime by Truck or Rail
Ii ~ 6.00
011.
(,)-
:I:
I-VJ
:I: 0
~ ~ 4.00
II::
I&.Z
~O
oJ I-
~ ~ 2.00
o
~S10NE x
\Jt-A
X
~ FREIGHT RATE FOR EACH
MATERIAL BASED ON
LOWEST COST METHOD
OF TRANSPORTATION
o
o
2S0
Figure 8. Comparison of Bulk Freight Rates
for Limestone, Calcined Limestone, and
Hydrated Lime
-325 mesh), for which the cost of grinding from
1/4- by O-inch size is estimated to be $0.32
per ton. The cost of grinding to finer parti-
cle size has also been est~mated (Fig. 9) (see
Appendix IV for basis and calculation).
Also included in Figure 9 are incremental
values above the base cost of grinding to 70%
36
minus 200 mesh. Therefore the effect of finer
grinding on overall injection cost can be ob-
tained by adding the appropriate amount to the
cost of limestone in Figure 14.
The grinding costs given are for material
with a Hardgrove grindability index of 50. In
tests at the TVA Colbert plant with the lime-
stone assumed as the sorbent in this study,
grindability was found to be 57 for the lime-
stone and 61 for the coal. It appears that,
in general, limestone should be no more diffi-
cult to grind than coal. However, further
data are needed on range of grindability among
limestones.
Curves relating particle size to cost of
limestone per stoichiometric equivalent of sul.
fur dioxide removed are given in Figure 10
The curve is derived from the relationship;
C = L + ar-n
br-m
where C = effective cost per ton of ground
limestone actually utilized
L = delivered cost of limestone
r = particle radius
The derivation of this equation and the basis
for the curves are given in Appendix IV.
For Tanaka's sorption data, Figure 10
-------
W
Z
o
.... J: W 0.80
(f)(f)z
Ww
::E::E~
..J&O(f)
NW
C)I")::E
~ '0:J 0.60
o~
z~z
(E'I1"0
(.!)La..""
0"""
~W~Q40
N-
t;u;~
O..JW
0<3:::e
..Jj::o
~ z~ 0.20
z .
w::e~
::Eoo
wcro
crLa..~
o -
z
~
. . -1'~-- ._--,,,,,. r '~'~~'''~I ~ ~ ..
RAYMOND RING- ROLL MILL I M,.. 0
35,000 LB./HR. LIMESTONE GRINDING RATE I w - Jot:
HARDGROVE! NDEX: 50 , :;::::t
POWER COST: 3MILLS/KW-HR Ii ~==
CAPITAL CHARGE: 11% OF INVESTMENT i13 ~w
MAINTENANCE: 11°/0 OF INVESTMENT ~ w z
0,90 :J NO
I -r-
Cj; 00 00i
~ ~ fj~
CW::::J
.24
0.10 (E:z: z
t~UO
I OC 1--=
. ~ ::)"
I ~...
1 t::~
'1#
-------
C= L+1.48r-O.5
or-It
L:: $ 2.00/TON, DEL. COST OF UNGROUND LIMESTONE
r '" PARTICLE RADIUS
L48r-O.5::GRiNDING COST, $/TON. FROM BONO EQUATION
or - n = WT. FRACTION OF LI MESTONE REACTED
0-0=0.272 i n=0.33USHIHARA DATA)
0-0 = 0.222; n = 0.137 (TANAKA DATA)
20
z
o
I-
.;;;. 18
a
w
N
...J
i= 16
::>
w
z
o
I-
~ 14
~
...J
I.L.
o
I- 12
en
o
(.)
10
8
100 70
20
0.1
10 1 5 3 2 I
AVERAGE PARTICLE SIZE. MICRONS
(50 PER CENT MINUS INDICATED SIZE)
50
30
0.1
0.5
0.3
0.2
Figure 10.
Effect of Particle Size on Cost of Limestone Utilized
quoted by vendors, which include profit and
sales cost. As has been noted, long-term con-
tracting for large amounts would likely re-
duce vendors' prices considerably.
The values given in Figure 11 are for a
battery limits1 calcination plant, assuming
raw limestone delivered to the calcination
unit. Therefore, subsequent estimates of over-
all injection cost of limestone can be con-
verted to a calcined limestone basis merely by
adding the appropriate calcination cost. It
has been assumed that the cost of grinding the
calcined limestone is the same as for raw lime-
stone; actually, the calcine should be some-
what easier to grind, but the saving is not
likely to be a significant one.
Hydration: Operating costs for manufac-
turing hydrated lime are also given in Figure
1 Erected unit only; services to unit, storage,
and all accessory items supplied by owner.
38
11. These can be used in the same way as the
calcination cost figures to get total injec-
tion cost for hydrated lime, by combining raw
lime~tone injection cost with the cost of man-
ufacturing hydrated lime. This appears con-
servative since elimination of grinding, which
presumably would not be needed for hydrated
lime, should give some saving; the greater
difficulty in handling hydrated lime might off.
set this to some extent, however, so that it
seems best to take no credit for eliminating
grinding.
Investment
The overall capital cost for a 200-mega-
watt limestone injection system--including
sorbent receiving and storage, grinding, in-
jection, separation from gas, and disposal--
is shown in Table VI. A comparable estimate
(Table VII) is given for a 1000-megawatt unit;
this is generally applicable either to a sin-
-------
20.00
ASIS:
LIMESTONE RAW MATERAL,95°/oC03
CALCINED LIMESTONE ,98% CoO
HYDRATED LlME,99% Co(OH)2
8000HR./YR. OPERATION
SORBENT REQUIREMENT
POWER
% STOICH. PLANT
CoO SIZE,MW
-
~
U
..J
Ocr
~-
~ffiO
u ~ ~ 12.00
~~~
~~~
~~ ~ 8.00
;:)wo(/).
z>
cr-
2(/)
d
)(
~
°/0 S
IN COAL
1: 2.0
m 2.0
:nz: 3.5
Jr 3.5
4.00
100
200
200
200
200
200
200
1000
HYDRATED LIME
CALCINED LIMESTONE Jr
Figure 11. Manufacturing Cost Vs. Plant
Capacity: Calcined Limestone and Hydrated
-Lime (Exclusive of Raw Material Cost)
Table VI. Summary of Estimated Fixed
Investment Requirements: Dry
Limestone Injection Systema_-
200-Megawatt Power Plant
Yard improvement!
Road and general yard modifications
Limestone storsge snd handling fscilities
Concrete foundations and conveyor tunnel
Receiving hopper, storage silo, conveyor
supports and bridges, and powerhouse
storage s 110
Powerhouse structura1 steel rev1 SiODS
Elevsting snd conveying in powerhouse
Pulverizing snd air system equipment
Injection equipment (dense phase)
Electroststic precipitators for incremental
dust loeding
Weste disposel equipment
Fly ash handling end disposal facHi ties
Waete disposal land requirements
60 acres (10-yr. requirement)
Control equipment
nectrical work
Painting and miscellaneous
Construction facHi tiee
Total direct cost
Engineering design
Contractor tees and overhead
Contingency allowance
Total project investment
Investment, $
25,')()()
80,000
71,000
} ,000
76,000
162,000
50,000
2}4,000
172,000
12 ,000
85,000
110,000
20,000
50,000
1,150,000
115,000
170,000
115,000
1,550,000
a Basis:
200""'.
}. 5 ~ sulfur in coal
2oo~ stoichiometric calcium oxide
Other bases 88 outlined under "Study Assumptions
and Design Criteria"
Table VII. Summary of Estimated Fixed
Investment Requirements: Dry Lime-
stone Injection Systema_-1000-
Megawatt Power Plant
Investment, $
Yard improvements
Road and general yard modifications
Limestone storage and hendling facilities
Concrete foundations and conveyor tunnel
Receiving hopper, storage s110, conveyor
supports end bridges, and powerhouse
storage silo
Powerhouse structural steel revisions
Elevating and conveying in powerhouse
Pulverizing and eir system equipment
Injection equipment (dense phase)
Electroststic precipitators for incremental
d11St loading
Waste disposal equipment
Fly ash handling and disposal facHities
Weste disposal land requirements
}OO acreS (lO-yr. requirement)
Control equipment
Electrical work
Painting and miscellaneous
Construction facilities
50,000
180,000
200,000
10,000
200,000
450,000
140,000
725 ,000
4S1J,000
60,000
lSIJ , 000
250,000
40,000
100,000
~ ,085 ,000
Total direct cost
Engineering de. ign
Contractor fees Bnd overhead
Contingency allowance
215,000
}40,000
--2!Q., 000
Total proj ect investment
},950,000
8. Basis:
1000 !I!W.
3.5 sulfur in coal
2oo~ stoichiometric calcium oxide
Other bases as outlined under "Study AS8\m1pt1ons
and Design Criteriall
gle 1000-megawatt unit or to a group of small-
er boilers totaling 1000 megawatts served by
a central grinding system. These estimates
are based on the amount of limestone required
for 3.5% sulfur in the coal and twice the stoi-
chiometric proportion. The effects of varying
the amount of lime injected in accord with sul-
fur content and stoichiometric ratio are shown
in Figure 12. The curve for limestone injec-
tion covers the complete injection system--re-
ceiving) storing) grinding) injecting) removal
in electrostatic precipitator) and disposal.
Included also) for convenience) are curves for
the battery limits investments fQr calcination
and hydration. Tbtal investment for hydrate
injection) for example) can be obtained by add-
ing costs for calcination) hydration) and lime-
stone injection.
Operating Cost
Operating cost for a 200-megawatt unit)
with limestone delivered at $2.05 per ton, is
given in Figure 13 as a function of limestone
consumption. Overall cost for injecting cal-
cined limestone or hydrated lime is shown also.
Detailed estimates for the 200-megawatt
unit at each of four levels of limestone con-
sumption are given in Appendix IV.
For convenience) a curve relating deliver-
39
-------
10.00
..
:>-
..-
u
it 6.00
z
I
n
m
]Jl
PLANT SIZE, 200 MW
8000 HRS. /YEAR OPERATION
@/o S IN COAL °/0 STOICH. CoO
2.0 100
3.5 100
2.0 200
:3.5 200
HYORATiON PLANT
.
Dr
~
t
:m
Q.
o
10 20 30 40 50 60 70
SORBENT CONSUMPTION. THOUSAND TONS CoO PER YEAR
80
Figure 12.
Investment for Limestone Injection, Calcination, and Hydration
ed limestone cost to overall operating cost has
also been prepared (Fig. 14). The effect of
unit size is shown by estimates for a 1000-
megawatt unit (Appendix IV); estimates are in-
cluded for limestone, calcined limestone, and
hydrated lime injection. The results are
summarized in Table VIII.
Finally, the effect of intermittent as
40
compared with continuous operation is shown in
Figure 15. For the base level chosen for in-
termittent operation, ten 3-day periods per
year; annual cost for a 200-megawatt unit (with
3.5% S in coal and twice stoichiometric) is
about $220,000 as compared with $625,000 for
continuous operation.
-------
0.936 2.50
o 0
L&J L&J
!;x Z
a:: a::
L&J ::>
z 0.749 «12.00
L&J ...I
C)
Z
Z
-------
e 5.00
iZ
er
;::)
m
...J
g 4.00
o
"-
o
z
~ 3.00
"-
-0-
l-
V>
8 2.00
CI
z
~
er
~ 1.00
o
...J
i!
o
I-
PLANT SIZE,200MW
SULFUR CONTENT OF COAL,3.5°/.
STOICHIOMETRIC RATIO, 2.0
INJECTION TIME, 8000 HRS./YR.
COAL BURNED, 600,000 TONSI YR.
o
o
5.00
Figure 14.
Effect of Sorbent Cost on Total
Operating Cost
Table VIII.
Effect of Inj~ction System Size
on Injection Costa
Sorbent
Overall cost/ton of coal, $
200 mw. 1000 mw.
Limestone
Calcined limestone
Hydrated lime
1.050
1. 732
2.100
0.863
1.305
1.540
a Basis: $2.05/ton of limestone; particle
size) 70% -200 mesh; 3.5% sulfur in coal;
200% of stoichiometric.
42
In" 700
o
ov>
CI ~ 600
~:J
!:to 500
ero
1&1,,-
2>0400
...JV>
:5 ~ 300
zex
zV>
ex 5 200
...Jx
-------
RESEARCH AND DEVELOPMENT NEEDED
It is obvious from the foregoing discus-
sion (particularly under "Status of Limestone
Injection") that much remains to be learned
about sorption by limestone, both in regard to
removal efficiency and operating difficulties
in simple injection and to new departures that
might give better performance.
Sorbent Efficiency
Further data are needed on the several pa-
rameters involving properties of the sorbent
and the kinetics of sorption: (1) limestone
versus dolomite, (2) intrinsic properties that
affect porosity after calcination, (3) raw
stone versus oxide versus hydrate, (4) parti-
cle size, (5) type and amount of iron compounds
present, (6) addition of promoters, (7) con-
centration of sulfur dioxide in the gas, and
(8) amount of excess air in the gas. These
are all being studied in small-scale and pilot
plant programs under way. In addition, it may
be possible to get full-scale data on some of
the variables in the NCAPC-TVA full-scale test
program.
Fine Grinding
There is substantial indication that grind-
ing to very fine particle size (on the order of
10 microns or smaller) will increase sorption
enough to more than offset the additional cost
of grinding to such fine size. Grinding to
this degree of fineness the large tonnages of
limestone required would be a new departure in
grinding practice. A study of fine-grinding
techniques is needed to minimize the cost of
operation on a large scale; fine grinding in
current practice is applied mainly to low-vol
ume materials that sell at a relatively high
price.
Emphasis in the study should
wet grinding, with dispersants to
glomeration, and injection of the
slurry into the boiler.
be placed on
prevent ag-
resulting
Method and Place of Injection
As has been discussed, good distribution
of the injected limestone presumably can be ob-
tained if the limestone is injected with a
large body of gas--such as the combustion air
(injection with the coal) in pulverj.zed-fuel
furnaces or the tempering or recirculating gas
in furnaces fitted for gas recirculation. In-
jection with the fuel, however, may not be
feasible. As noted earlier; there are con-
flicting data on the question. Further data
are needed to establish whether there is any
loss of sorption efficiency from dead-burning
and, if so, whether it is significant enough
to outweigh the advantages that accrue from
injecting with the fuel.
The problem in injection with the fuel
appears to be one of particle size. As point-
ed out earlier, some of the tests have shown
as good results as in tests of injection above
the flame; however; the sorption still was not
good in the sense of getting good limestone
utilization. Therefore, finer particle size
should be tested as a means of improving sorp-
tion (size was about 200 mesh in the tests).
But this would make the limestone more subject
to dead-burning.
Tests with finely ground limestone (325
mesh) have been carried out recently at the TVA
Colbert plant but analyses of the products are
not yet available. Other tests are also under
way or planned--injection with oil by the
Florida Power and Light Company, moving-grate
stoker studies of limestone-coal feeds by the
Peabody Coal Company, and pilot plant tests by
Band W.
If the sorbent must be injected above the
burners, the problem of uniform distribution
(in proportion to cross-sectional mass gas flow)
becomes paramount. Developing and testing a
method for obtaining good distribution will be
one of the major objectives in the full-scale
program.
A further problem is the wide variation
in boiler type. Since the NCAPC-TVA program
will test only two types, a study will be need-
ed on extrapolation of the findings to other
types.
Slagging and Erosion
Conflicting results have been reported al-
so for effect of limestone injection on slag
formation and boiler erosion. Long-term tests
such as planned in the full-scale project
should be helpful in resolving this question.
As many combinations of coal type and limestone
type as possible should be tested in order to
cover the full range of slag characteristics to
be expected. Small-scale tests on the problem
are being made by Band W.
Reaction between limestone and fly ash may
also have an adverse effect on sorption. Driv-
ing the limestone particles at right angles in-
to the curtain of fly ash particles probably
gives adequate opportunity for impingement, and
it has been noted in TVA studies that many of
the recovered particles contained both lime and
ash. Reidl has treated this problem; further
l Reid, W. T. In Final Report, "Fundamental
Study of Sulfur Fixation by Lime and Magnesia,"
Battelle Memorial Institute, June 30, 1966.
43
-------
study is needed on a pilot plant or plant scale.
Increase in Retention Time
Probably the main drawback to the lime-
stone-injection method is the very short reac-
tion time available in a power boiler. In only
a second or so (for most boilers)) the steps of
limestone calcination) diffusion of sulfur ox-
ide into the particle interior) and reaction of
calcium oxide and sulfur oxide must take place
(plus oxidation at some point). It would be
quite desirable to increase the time available
for these operations.
It has been pointed out that increasing
retention beyond a certain length of time may
do little good) because sorption becomes ex-
ceedingly slow after the calcium sulfate shell
on the particle reaches a certain limiting
thickness. In small-scale) fixed-bed tests the
sorption rate rapidly reaches a plateau after
which sorption is so slow as to be of little
significance. Relatively large particles were
used in these tests) however) for which the
shell of "critical thickness" made up less than
half of the volume. Therefore it is essential
to grind the limestone to a particle size small
enough that the shell formed makes up a major
portion of the particle volume. This is essen-
tial no matter what the retention time may be.
Having done this) then an increase in retention
time should help in getting as near as possible
to any limiting shell thickness.
There are four main ways by which reten-
tion time might be increased in standard boil-
ers: (1) deposition of sorbent on boiler sur-
faces) either transitory or) if held by sticky
slag) semipermanent (until removed by sootblow-
ing)) (2) impingement on barrier surfaces with
accompanying rebound) (3) countercurrent injec-
tion) and (4) sorbent recycling. There are
other ways of increasing retention time) of
course) if we include special reactors or types
of boilers. For example) NCAPC is presently
funding two projects (see p. 2) in which the
fluidized-bed technique is being tested as a
means of increasing retention time.l The cy-
cling) entrained reactors used in the Still
process2 also give extended retention. The
present study) however) is limited to those
methods that might be used in existing types
of boilers with no more than minor alterations.
Sorption of sulfur dioxide in material de-
posited on boiler walls and tubes is one of the
least understood factors in limestone injec-
tion. Ishihara found) in tests in an oil-burn-
ing boiler (Appendix I)) that sulfur dioxide
l Good results (90% Cao utilization) have been
reported recently in one of these projects.
2 Brocke) W. Report for Third Limestone Sympo-
sium) Clearwater) Florida) December 4-8) 1967.
44
removal increased gradually after injection was
begun) and reached a constant value after 2
hours. Dust deposits taken from the boiler
walls after the test contained 80 to over 90%
calcium sulfate at points where the boiler tem-
perature was between 900° and 1000° C. At
points farther along the boiler (temperature of
200°-600° C.) the deposits contained only 40%
calcium sulfate) indicating that part of the
sorbent was deposited on the walls at the be-
ginning of the injection and that it sorbed
sulfur dioxide in accordance with the tempera-
ture driving force.
Goldschmidt's tests in a 22-megawatt oil-
fired boiler (Appendix I) confirmed this. With
coal firing) however) he found that the depos-
its did little sorbing) presumably because the
deposited limestone quickly became covered over
or reacted with molten slag. This was advanced
as the reason for the much better sorption when
the boiler was fired with oil.
Further data on the effect of deposits
probably can be obtained in the full-scale
testing program. In addition) it may be worth-
while to consider pilot plant tests of impinge-
ment barriers) inserted in the boiler) on
which the sorbent might deposit; a major point
to be determined would be how far the barriers
would have to be placed away from the burners
to avoid the effect noted by Goldschmidt.
An extension of the impingement barrier
concept has been proposed by A. J. Teller (The
Cooper Union for the Advancement of Science
and Art).3 The barriers would be ribbons of
metal under sufficient tension to produce vi
bration in the gas stream. The vibration might
help shake reacted particles off and make way
for new ones. Finding a suitable construction
material) of course) would be a major problem
in using any type of barrier.
Ideally) any barriers used should be plac-
ed in the boiler at a point where the tempera-
ture is optimum for sorption in calcined lime-
stone at steady temperature) namely) 1700° to
1800° F. For some boilers at least) this would
be in the convection pass where occlusion in
molten slag might not be a problem.
Method 2) impingement and rebound) would
also involve use of barriers but they might be
sized and placed differently. The method is
probably more applicable to the larger parti-
cles in a particular grind of limestone than
to the finer ones because the latter would re-
quire more pressure drop for impingement and
are more likely to stick than to rebound.
It might be said that this type of reten-
tion would help because it would keep the
larger particles in the boiler longer than the
finer ones (say) 80- vs. 5-micron sizes) and
3 Teller) A. J.
Private communication (19~
-------
thus provide the longer period required for
complete calcination and sulfur oxide diffusion
to the interior of the large particles. Coun-
tering this is the argument that only a certain
thickness of shell will be formed regardless of
retention time (in a practical range) and that
large particles therefore need little or no
more time than fine ones to do all they are go-
ing to do. If the reaction layer thickness is
limited) then there may be no need to calcine
the particle core and incomplete calcination
may therefore be acceptable (unless continuing
C02 evolution interferes with S02 sorption).
In the discussion on grinding in Appendix IV)
however) it is brought out that plant tests
have indicated a thicker shell on large parti
cles than on small ones (0.75- and 0.143-micron
shells on 40- and 4-micron particles) respec-
tively). The situation is not yet clear.
It should be emphasized that putting any-
thing in the way of barriers in the boiler is
very questionable from the standpoint of boil-
er operation. Such a departure should be con-
sidered only if tests indicated a major im-
provement in sorption. Then the effect on
boiler operation could be evaluated to deter-
mine the overall balance of advantage and dis-
advantage.
Method 3) countercurrent injection) was
discussed in an earlier section. Either with
wall-mounted injectors or nozzles spaced over
the boiler cross section) it should be possible
to inject downward against the main gas flow
and get some increase in retention time by in-
ducing an initial period of cn1mtercurrent
flow. Whether or not a significant increase
in retention could be obtained by this method)
und2r the practical limitatious of injection
fluid velocity) is not known. Calculations
and possibly experiments should be made by
someone with experience in fluid dynamics to
evaluate the possibilities.
If it develops that the sorbent could be
driven a significant distance countercurrently)
there might be some question then as to how
this would affect the optimum injection point.
If we assume) as Zentgraf's data indicate) that
limestone has sufficient time to calcine be-
tween 2700° and 2100° F.) and that there is no
advantage in having calcined limestone in the
system above 2100° F.) then nothing is gained
by increasing retention time of limestone above
2100° F. What is needed is longer retention
time of calcined limestone at or near the tem-
perature best for sorption) 1700° to 1800° F.
Hence countercurrent injection may have
more application to calcined or hydrated lime.
If these materials were injected from a point
in the convection pass where the temperature
was) say) 1600° F.) in a backward and downward
direction toward the 2100° F. level (which for
the Colbert No.2 boiler in Fig. 1 lies in the
boiler throat) the sorbent would pass through
the optimum temperature zone twice and the
overall retention time in the "effective tem-
perature" zone would be increased considerably.
The main question with this is whether or
not the sorbent would pick up any sulfur di
oxide during the period of countercurrent flow.
Since it would be "encased" in the jet of in-
jection fluid) there might be little contact
with the main body of gas until the jet broke
up.
Therefore there might be little increase
in "effective" retention time even if the ac-
tual time were increased somewhat by counter-
current injection. There might be an inci-
dental advantage) however) from preheating of
the particles during the countercurrent flow.
Moreover) the fact that the sorbent would be
at zero velocity when released into the gas
stream at the point of jet breakup may be help-
ful; in contrast) a jet introduced horizontally
would bend upward and the particles therefore
would be moving cocurrently when the jet break-
up released them into the main gas stream.
Thus the initial zero velocity obtained by
countercurrent injection might actually in-
crease effective retention time.
From this it appears that countercurrent
injection is a complicated matter and that a
considerable amount of further study will be
required for evaluation. A thorough fluid
dynamics study) taking into account the chem-
ical data available) is needed.
Method 4) recycling of sorbent) is a com-
mon expedient in other chemical processes in
which retention time or equilibrium consider-
ations prevent adequate reactant conversion
in a single pass. There are major problems)
however) in applying recycling to limestone
injection. First) the separation of reaction
product from unused reactant before recycling
is difficult in limestone injection) particu-
larly in coal-burning plants where the lime
is collected as a mixture with fly ash. Sec-
ondly) if the total mass is recycied merely
as a means of getting more retention time)
the shell of calcium sulfate formed in the
first pass interferes with sorption in the
second pass.
For oil-burning plants) these problems
are not as severe. Ishihara has proposed a
method in which limestone is injected into the
boiler) recovered in dust collectors) hydrated)
and reinjected--thus giving two passes through
the boiler. The main advantage of this ap-
proach is that the hydration should break down
the particles and) as a result of exposing
fresh surfaces) eliminate the adverse effect
of the sulfate shell in the second pass. More-
over) there is some evidence that calcination
45
-------
followed by hydration and dehydration increases
porosity. An economic study relating cost of
hydration and recycling to the increase in
sorption required to justify the cost would be
helpful in evaluating this method.
Ishihara proposed hydration with only the
stoichiometric amount of water so as to get a
dry powder for reinjection. It might be pre-
ferable, from the handling standpoint, to hy-
drate with excess water and return the lime to
the boiler as a slurry. Moreover, Zentgraf's
work indicates that the initial injection of
limestone should be at 2700° F. or higher; but
that the reinjection of the hydrated lime
should be at about 2100° F. Hence two injec-
tion systems would be necessary.
Steam might also be used as the carrier
for injecting recycled calcium oxide, there-
by perhaps hydrating during travel with the
jet and dehydrating after the jet broke up.
Another variation could be reinjection of
the hydrated lime into a Still-type system
operating on relatively cool gas. The in-
creased retention time in such a system might
result in high sorbent utilization.
A further problem is that in operation
of any recycle process the recycle would con-
tain second-pass as well as first-pass mate-
rial. The system could only be kept in bal-
ance by discarding part of the total mass from
the dust collector, in which case part of the
first-pass material would be discarded, much
as in an ammonia synthesis loop where a por-
tion of the incoming synthesis gas is lost
in bleeding off part of the recirculating
loop gas to prevent buildup of inert methane
and argon in the loop. The amount of valuable
gas (or first-pass calcined limestone) lost
depends on the ratio of recycled material to
fresh material in the loop. In the Ishihara
method, the ratio of recycle to fresh sorbent
in the total mass injected could be kept high
to reduce the amount of first-pass sorbent
passing directly to discard. However, mate-
rials handling and dust collection could be-
come a problem and the relatively large amount
of solid material in the gas stream could
erode the boiler and interfere with heat trans-
fer. If the total mass injected were held to
twice the amount of raw limestone fed (one
part limestone and one part recycle), then
half of the first-pass material would be dis-
carded and the remaining half would be the
only sorbent benefiting from recycle. To in-
crease this to 75% recycle of the original
material, three weights of recycle to one of
fresh limestone would be required, which would
throw a very heavy load on the solids-transfer
and dust-removal systems.
For coal-burning plants, the presence of
fly ash would be a major complication. Work
46
is under way (under a NCAPC contract with West
Virginia University) on beneficiating to sepa-
rate ash from calcine but the project is in
the preliminary stages. Small-scale tests at
TVA have indicated that separation would be
difficult. If the ash-lime mixture were hy-
drated perhaps the very fine hydrate particles
could be separated. Even so, however, the re-
cycle problem described above would give dif-
ficulty.
Internal recycling in the boiler system
has been proposed. Schemes have been suggest-
ed in which the size of the limestone parti-
cles injected would be large enough to allow
rough separation from the fly ash by insertion
of a "knock-out" device in the boiler, either
in the convection pass at the top of the slope
(see Fig. 1) or just before the air heater.
The separated sorbent would be recycled, by
"blowing" back down the slope of the convec-
tion pass or by removal and reinjection if re-
covered at the air heater. Both of these have
the drawback that, if the limiting shell thick-
ness theory is correct, recycling may not ac-
complish very much. The current work on flui-
dized contactors (see p. 2) should throw light
on this point.
Another type of recycling is regeneration
of the product calcium or magnesium sulfate to
give oxide or hydrate for recycling; sulfur ox-
ide or sulfur would be a coproduct that would
help pay for the regeneration. This approach
is being studied by Esso Research and Engineer-
ing under contract with NCAPC. Preliminary
studies indicate that probably only magnesium
sulfate regeneration should be attempted, in
which case any calcium compounds in the sor-
bent would ride around the circuit as inerts.
Obvious problems in the method are (1) getting
fly ash separated and (2) using magnesium ox-
ide as the sorbent, since sorption rate at the
thermodynamically optimum sorption temperature
may not be high enough.
Further exploration of the recycling ap-
proach appears warranted but at the moment the
method does not seem promising for coal-fired
boilers.
There are several other possibilities for
increasing retention time. Induction of large,
stable eddies has been suggested, for example,
as well as the high-velocity jets mentioned in
the next section. All of these require analy-
sis from the fluid dynamics standpoint before
they can be evaluated. Because of the great
need to improve sorption efficiency, such
analyses should be made.
Turbulence Promotion
Increase in turbulence, as a possible
means of promoting mass transfer (see p. 17),
would be obtained incidentally by the impinge-
-------
ment barriers mentioned previously. The colli-
sion of particles with the barrier and the re-
sulting change in direction would produce rel-
ative motion of particle and gas. If turbu-
lence were found to be a major factor in sorp-
tion, however, it might be better to design
the barriers for maximum turbulence rather th~n
for maximum particle retention time. Perhaps
there would be no difference; no attempt has
been made in this study to analyze the fluid
dynamics factors involved. Such an analysis
should be made.
Introduction of recirculated gas through
high-velocity jets in the 1700° to 1800° F.
zone possibly could give some benefit by in-
ducing relative motion between particles and
gas and perhaps at the same time increase re-
tention time by slowing the particles down.
Jets might thus take the place of impingement
barriers, which probably would be an easier
course from the construction standpoint.
Dust Collection
One of the major problems in use of lime-
stone in the dry state is removal of the large
mass of solids from the gas stream after pas-
sage through the boiler. Dust-removal equip-
ment in existing plants very likely would be
inadequate, both because of the additional
amount of solids to be removed and because
of the high resistivity of the limestone.
In tests at Detroit Edison, the precipi-
tator lost efficiency when limestone was in-
jected (Appendix I). In contrast, Zentgraf
did not encounter an efficiency loss in his
tests. It has been said that precipitators in
Germany are generally designed for relatively
low sulfur trioxide content in the gas (be-
cause of low sulfur content in the coal) and
high lime in the ash, and therefore are less
subject to efficiency loss from limestone in-
jection than are U. S. precipitators, which
are designed for high sulfur trioxide and low
lime. Conversely, it has been argued that the
calcium sulfate shell formed on the limestone
reduces its resistivity and makes it amenable
to collection in electrostatic precipitators.
The situation is not clear. Small-scale
testing is being carried out jointly by TVA,
Research-Cottrell, and Band W to determine
resistivity of reacted limestone. Large-scale
tests are needed. A limited study of the prob-
lem may be possible in conjunction with the
NCAPC-TVA plant test program.
Use of Spent Sorbent
The mixture of calcium sulfate, unreacted
calcium hydroxide, and fly ash resulting from
use of limestone injection produces a major
disposal problem. Research is needed to de-
velop some method by which the material can be
used profitably rather than piled up in un-
sightly mounds. Such research is being car-
ried out by West Virginia University under
contract with NCAPC.
47
-------
CONCLUSIONS AND RECOMMENDATIONS
The relative simplicity of limestone in-
jection as a means of removing sulfur oxides
from power plant stack gas is a very attrac-
tive feature. The boiler serves as the reac-
tor, no complicated regeneration system is
needed, and no reheating of stack gas is re-
quired.
Low degree of limestone reaction, how-
ever; is a major drawback. Unless new tech-
nology for improving sorption is developed, 25
to 30% utilization of the injected limestone
is probably the highest attainable. Excess
limestone can be used to increase sulfur diox-
ide removal, but adverse effect on power plant
operation and cost of removing the increased
dust burden from the gas and disposing of it
then become major disadvantages.
The cost of limestone injection, both for
investment and operation, depends mainly on
the amount of limestone used and the size of
installation. Investment for a 200-megawatt
unit, for example, assuming 3.5% sulfur in the
coal, is $5.90 per kilowatt of capacity for
stoichiometric injection of limestone and
$7.75 for 200% of stoichiometric. Correspon-
ding operating costs, assuming $2.05 per ton
of limestone, are 0.24 and 0.39 mill per kilo-
watt-hour generated. If the injection instal-
lation is increased to a lOOO-megawatt size,
investment (for 200% stoichiometric) is re-
duced to $3.95 per kilowatt and operating cost
to 0.325 mill.
Hence the investment and operating costs
for limestone injection are about 3 and 10%,
respectively, of normal power plant costs.
The main items in investment are the in-
creased costs of dust-collection and waste-
disposal equipment, representing about 25 and
20% of the total, respectively, for a 200-
megawatt unit (3.5% Sand 200% of stoichiomet-
ric limestone). For the operating cost, lime-
stone (at $2.05/ton delivered) accounts for
about 45% of the total, operating costs make
up 24%, and capital charges 31%.
The main opportunity for improving the
process is to increase limestone utilization.
Leading possibilities for doing this are as
follows:
1. Calcine or calcine-hydrate the lime-
stone before injection to increase poros-
ity and perhaps retention time. This in-
creases cost significantly, however; and
some of the work in Germany indicates that
raw limestone will sorb as well as the more
costly materials if injected at the optimum
point.
2. Increase retention time. Several meth-
ods have been proposed for increasing sor-
48
bent retention time in the boiler- Most of
these appear questionable but they should
be investigated further.
3. Increase mass transfer by developing
relative movement between particles and
gas. This is also somewhat dubious but
worth investigating.
4. Catalyze oxidation by selecting lime-
stones high in iron or adding a catalyst.
This is a promising approach that should be
explored further. Problems expected are
limited availability of high-iron limestone
at economical locations and difficulty in
incorporating catalyst intimately into the
limestone.
5. Improve porosity after calcination by
adding sodium salts to the limestone.
Tests should be made.
6. Reduce sorbent to extremely fine parti-
cle size. This is probably the most promis-
ing of the various approaches; although in-
crease in sorption efficiency does not ap-
pear to be proportional to the surface area
increase resulting from the fine grinding,
there is still evidence that the improvement
will more than offset the additional grind-
ing cost. A thorough study of fine grind-
ing economics should be made, with emphasis
on wet grinding followed by slurry injec-
tion. Hydrated lime, reduced to submicron
size by pressure hydration followed by sud-
den pressure release, should also be con-
sidered.
Work is also under way on regeneration of
the product calcium or magnesium sulfate,
which of course would increase utilization by
recycling much of the sorbent.
Further work is also needed on basic con-
siderations such as (1) the intrinsic proper-
ties of limestone that affect porosity after
calcination, (2) the relative effectiveness of
dolomitic limestone, and (3) ways of adjusting
product resistivity as a means of improving
precipitator operation. These points are being
studied in various programs under way.
The forthcoming NCAPC-TVA tests in opera-
ting boilers should afford an opportunity for
gathering much of the data needed, particular-
ly on the several questions regarding effect
of limestone on boiler operation. The follow-
ing major objectives are suggested for these
tests.
1. Test selected limestones for sorption
efficiency and slagging problems.
2. Develop injection system that gives the
proper sorbent distribution quickly.
3. Determine point and method of injection
that gives optimum initial temperature lev-
-------
el for limestone, calcined limestone, and
hydrated lime.
4. Gather data on benefit of small parti-
cle size.
5. Observe effects on power plant opera-
tion, including flame stability, heat trans-
fer, slagging, corrosion, erosion, dust col-
lector performance, problems in slaking
calcium oxide, and water pollution.
6. Determine effect of power plant load
factor on sorption.
7. Gather data to refine cost estimates.
8. Determine effect of recycling sorbent
in the recirculating gas (one of the test
boilers is of the recirculated-gas type).
49
-------
APPENDIX I
LIMESTONE SORPTION STUDIES BY VARIOUS ORGANIZATIONS
Since much of the work on limestone injec-
tion quoted in the main body of this report
has not been published, summaries of the vari-
ous investigations will be given as background
for the Qssumptions and conclusions of the de-
sign study. Only those investigations of par-
ticular pertinence to this report are summa-
rized; other important work has also been car-
ried out but complete coverage has not been
attempted.
Tennessee Valley Authority (Muscle Shoa~
Alabama) 1
Limestone injection was tested at TVA as
far back as early 1955. A small coal-burning
pilot plant was used to supply stack gas for
the first tests. The unit was designed pri
marily to supply gas for an ammonia-scrubbing
process, however, and did not have the range
of temperature and retention time in the high-
temperature region that is needed for study-
ing limestone injection. Therefore, emphasis
was shifted to small-scale tests with simu-
lated stack gas or air-sulfur dioxide mixtures.
Materials tested were ground to pass a
200-mesh screen and were dried at 220° F. for
about 24 hours. Analyses of the materials
were as follows:
Limestone
Pelham, Rockwood,
Ala. Ala .
Dolomi te
Unknown
Chemical
analysis, %
GaO
Si02
Fe 20S
Al20s
MgO
Ignition
loss 41.3 43.5
H20 0.2 0.1
In all the tests the gas rates were adjusted
so that the inlet gas contained 0.29% sulfur
dioxide and the retention time in the heated
zone of the reaction tube was 1 second. The
exit gas temperatures from the heated zone
were in the range 1150° to 1370° F. With each
material, tests were run with several propor-
tions of limestone and with air as the carrier
gas.
50.9
4.5
0.9
0.3
1.7
55.0
0.5
0.1
0.1
0.7
21
30
Results of these tests are shown in
Figure 1-1. The data show that the two lime-
stones were about equally reactive but that
1 Unpublished progress reports.
100
90
80
70
~
;. 60
o
VI
IL
o 110
..J
~ 40
ILl
0::
o
LIMESTONE
o PELHAM, ALABAMA
6 ROCKWOOD,ALABAMA
DOLOMITE
C SOURCE UNKNOWN
TEST CONDITIONS
EXIT TEMPERATURE,
1150. TO 1370.F
RETENTION TIME, 3 SEC.
S02 IN INLET GAS, 0.29 %
30
20
10
o
o 100 200 300 400 1100 600 700
PROPORTION OF LIMESTONE, ". OF STOICHIOMETRIC
Figure 1-1. Effect of Limestone Proportion
on Removal of Sulfur Dioxide from
Air-Sulfur Dioxide Mixture
the dolomite was appreciably less effective.
The percent of sulfur dioxide removed from the
gas increased from 20 to 70 as the amount of
limestone fed was increased from 100 to 400%
of the stoichiometric proportion. About 90%
of the sulfur dioxide was removed when 600% of
stoichiometric was used. The proportion of
sulfur dioxide removed from the gas in the
tests with 100% of stoichiometric was of the
same order of magnitude as in the pilot plant
tests in which this proportion of limestone
was injected into the combustion chamber. In
the small-scale tests, no effect of tempera-
ture on the reaction could be detected in the
exit temperature range of 1150° to 1370° F.
These tests were made in a reaction tube
1 inch in diameter and 3 feet long with an up-
ward gas velocity of about 3 feet per second.
To make tests with longer contact times, the
length of the reaction tube was increased to
6 feet; the gas velocity was held constant.
A mixture of dry air and sulfur dioxide con-
51
-------
UNE AVERAGE GAS CONTACT
SYMBOL yo CARRIER GAS TEMPo, @F. TIME, SEC.
t AIR (DRY) 1645 0.8
{NO DATA POINTS SHOWN; CURVE FROM PREViOUS DATA}
o 2 AIR WRY) 1658 1.8
l:. :3 AIR WRY) 1425 1.8
o NZiCOZ,OZWRY} 1658 1.8
lIB] 4 N2,C02,O~dMOIST) 1658 1.8
8
1'0
fI!. 60
N
o
ro 50-
I.t.
o
~ 40
>
"
~
w
oc 30 ~
!
I
I
I
I
I /,
11/
10 11/
1,1'
20
COLUMBIA
U MESTONE
00
2
:3
o I
STOiCHOMETRIC RATIO
6
2
:3
4
5
Figure 1-2. Effects of Contact Time, Temperature, and Carrier Gas
Composition on Dry Limestone-Sulfur Dioxide Reaction
taining 0.29% sulfur dioxide was used as in
the previous tests. Tests were made with both
Columbia and Pelham limestones. Pelham analy-
ses are shown in the preceding tabulation;
Columbia limestone contained 45.5% calcium ox-
ide (no other analyses were made of this ma-
terial). Temperature traverses made through-
out the lengths of the two reaction tubes
showed that the length of the reaction zonel
was 28 inches in the short tube and 63 inches
in the long one: the respective contact times
were 0.8 and 1.8 seconds. The average tem-
peratures in the reaction zones were 1645° F.
for the short tube and 1658° F. for the long
one; this temperature difference probably was
1 Reaction zone is considered to be that por-
tion of the tube in which the gas tempera-
ture is 1200° F. or above.
52
not significant.
The results in Figure 1-2 show that the
increase in contact time from 0.8 to 1.8 sec-
onds decreased the requirements of both lime-
stones. Calculations based on the slopes of
lines 1 and 2 in the figure indicate that the
requirement of Columbia limestone was decreas-
ed from 25 to 13 pounds2 per pound of sulfur
removed and that of Pelham stone was decreased
from 18 to 14 pounds per pound of sulfur.
Using Columbia limestone and a contact
time of 1.8 seconds, a series of tests was
made at an average temperature of 1425° F. in
the reaction zone for comparison with the tests
at 1658° F. Calculations based on the slopes
2 The stoichiometric requirement is 3.1 pounds
of calcium carbonate per pound of sulfur re-
moved.
-------
of lines 2 and 3 of Figure 1-2 indicate that
this decrease in temperature increased the
limestone requirement from 13 to 30 pounds per
pound of sulfur removed.
Tests were then made with synthetic stack
gas. Pelham limestone) a temperature of about
16~00 F.) and a contact time of 1.8 seconds
were used) since these conditions gave the
best results in the tests made with the air-
sulfur dioxide mixture. The synthetic stack
gas was made by mixing pure nitrogen) carbon
dioxide) and air to give a gas containing
approximately 81% N2) 16% C02) and 3% O2; in
half the tests water vapor was added to the
gas to give a dew point comparable with that in
plant gas (about 12~0 F.). As in the tests
with air; sulfur dioxide concentration was
controlled at 0.29%.
Data from these tests (Fig. 1-2) show
that limestone requirement when using the syn-
thetic gas was 21 to 26 pounds per pound of
sulfur removed as compared with 14 pounds per
pound of sulfur when the air-sulfur dioxide
mixture was used. Limestone requirement was
apparently about the same regardless of wheth-
er or not the synthetic gas was humidified.
More recently (1967) work has been done
at TVA) under contract with NCAPC) on the ef-
fect of limestone physical properties on sorp-
tivity.
The impurities in limestone were found to
alter during calcination) the iron minerals
converting to hematite (Fe20S) and magnetite
(Fes04) and the clay) mica) and silica gel to
a calcium silicate glass. The glass makes the
calcine harder and less reactive) and also
ties up part of the calcium oxide.
Dolomite gave higher porosity after cal-
cination) presumably due to the alternate
sheet structure of magnesium carbonate and
calcium carbonate and the earlier decomposi-
tion of magnesium carbonate.
Reactivity of the calcine could be re-
lated to the volume of pores between 17.~- and
0.035-micron diameter. There was also good
correlation with density when pores larger
than 17.5 microns were excluded.
National Center for Air Pollution Controll
Tests in a fixed-bed sorption unit have
been made by the National Center for Air Pollu-
tion Control (NCAPC) at Cincinnati) Ohio. Ten
i Harrington) R. E.) Borgwardt) R. H.) and
Potter) A. E. "Reactivity of Selected Lime-
stones and D:Jlomites with Sulfur Dioxide."
Paper presented at the American Industrial
Hygiene Conference) Chicago) Illinois)
May 1-5) 1967.
limestones and dolomites were chosen to give
high and low levels of calcium) magnesium) and
iron. Most of the tests were made with minus
10- plus 28-mesh particles and with a flue gas
made by burning natural gas. Results were
measured in terms of time required for sorbent
saturation to the point that 20% of the sulfur
dioxide passed through ("breakthrough point").
In some of the tests) the limestone was
calcined before placing in the sorber and the
effects of calcination time and temperature
determined. The samples varied in regard to
optimum calcination temperature; some did
better at 1700° F. and others at 1800° F. At
1800°) a calcination time of 2 hours was opti-
mum for most of the limestones.
Calcining conditions) however) were much
less significant than intrinsic differences
between limestones. Under constant reaction
conditions) samples varied from 30 to 1~0 min-
utes in breakthrough time.
In tests on effect of reaction tempera-
ture) 1600° F. was better than lower levels in
practically all tests. Between 1600° and
1800° F.) results varied; about half the sam-
ples performed better at 1600° and the remain-
der at 1800° F.
No correlation could be found between
chemical composition and reactivity.
Reaction rate was very fast. Complete re-
moval of sulfur dioxide was obtained in a resi-
dence time of only about 0.02 second until the
breakthrough point was reached. Total amount
sorbed before breakthrough) however; was only
10 to 40% of the stoichiometric amount. It
was postulated that formation of an impervious
calcium sulfate shell impeded reaction in the
later stages of a run. Tests at smaller par-
ticle sizes confirmed this; in comparison with
20.1% utilization for minus 20 plus 28 mesh)
27.6% was obtained with minus 28 plus 35 mesh
but only 7.4% with minus 14 plus 20 mesh.
In kinetics studies) there was some gas-
film resistance at gas velocities below 3 feet
per second but none was indicated at higher
velocities. Other tests showed that the sorp-
tion is first order in regard to the partial
pressure of the sulfur dioxide. It was con-
cluded that the reaction rate (for the par-
ticular limestone studied) can be expressed by
the relationship:
1 d Nsos = K Cs02
W dt Nso~
where W is the weight of calcined limestone)
Nsos is the cumulative weight of sulfur triox-
ide sorbed by the limestone) and Cs02 is the
concentration of sulfur dioxide in the gas.
53
-------
Bergbau-Forschung (Essen, Germany)1,2,S
The Bergbau-Forschung organization in
Germany has studied limestone injection ex-
tensively. In small-scale tests, the kinetics
of decarbonation of limestone were first in-
vestigated. In subsequent tests of sulfur di-
oxide sorption by dolomitic limestones, re-
sults differed with various samples; in some
cases both magnesium carbonate and calcium
carbonate reacted, in some the magnesium car-
bonate sorbed preferentially, and in still
others only the calcium carbonate reacted.
Taese results were obtained in tests in which
the sorbent temperature was increased con-
tinuously throughout the test; such noniso-
thermal conditions were said to be closer to
those in the boiler than the isothermal oper-
ation usual in fixed-bed tests.
Other tests led to the following con-
clusions:
1. Calcination and sulfur dioxide sorption
take place simultaneously rather than se-
quentially under the condition of slow in-
crease in sorbent temperature used in the
tests. In actual injection, however, the
very rapid rate of heating may favor cal-
cination initially.
2. With slow heating, complete conversion
of the limestone to calcium sulfate is ob-
tained, whereas in actual injection only
50% conversion is obtainable under the best
circumstances.
It is assumed that under fast reaction
conditions, the reaction is inhibited by
the sulfate shell formed, but with the long
retention time in slow heating the sulfur
dioxide can permeate the limestone particle
completely.
3. Limestones vary widely in activity. The
following tabulation shows results of tests
with three limestones along with data on the
same limestones obtained by NCAPC workers.
Sample
BCR 1337
1352
1362
S02 reacted,
ncc/g. Cao
Bergbau-Forschung~ NCAPCD
234
160
95
131
93
29
a Nonisothermal conditions; 1.0%
S02 in gas.
b Cao reacted at 871° C. until 20%
breakthrough; 0.28% S02 in gas.
1 Kruel, M., and J-/.1ntgen, H. Report for Third
Limestone Symposium, Clearwater; Florida,
December 4-8, 1967.
2 Van Reek, K. H., and J-!1ntgen, H. Ibid.
S Kruel, M., and J-!1ntgen, H. Unpublished re-
port (1966).
54
In further work, a pilot plant designed
to study the influence of temperature, resi-
dence time, and amount of sorbent was built.
Stack gas was provided by combustion of coal
(1.5-1.8% S) in an underfeed stoker burner;
sulfur dioxide concentration was 0.11 to 0.14%
by volume. A side stream of gas was drawn
through a vertical, cast alloy reactor tube 100
millimeters in diameter and 6 meters long heat-
ed throughout its length. Sorbent was injected
through a steel tube at the bottom of the re-
actor where the combustion gases also entered.
Velocity in the reactor ranged from 1 to 4 me-
ters per second.
Tests were first made with hydrated dolo-
mite (composition, %: Cao, 44; MgO, 30; Si02,
3.5; Al20s plus Fe20S' 2.2); particle size was
90.6% minus 60 microns. The effects of temper-
ature, amount of sorbent, residence time, and
particle size on degree of sulfur dioxide re-
moval by hydrated dolomite were studied. With
stoichiometric addition, desulfurization ranged
from 20 to 45% as the temperature was varied
from 400° to 900° C.; 900° was the highest tem-
perature tested. With three times stoichio-
metric addition at 900°, 80% sulfur dioxide re-
moval was achieved. With particle sizes of
less than 60 microns sulfur dioxide removal was
about 20% higher than for particles larger than
60 microns. It was concluded that residence
times longer than 1.5 to 2.0 seconds do not in-
fluence degree of desulfurization. The effects
of temperature and amounts of sorbent on desul-
furization are shown in Figure 1-3 and the ef-
fect of particle size in Figure 1-4.
Tests were not reported for limestone.
Precipitated calcium carbonate gave very good
results at 800° C.--42 to 51% removal with
stoichiometric injection and 80 to 96% with
~
STOICHIOMETRIC RATIO
.
~ 60
o
JI
'"
II: 40
.,
o
en
20
o
300
Figure 1-3. Effect of Temperature and -'''.;;10unt
of Hydrated Dolomite on Sulfur
Dioxide Sorption
-------
100
80
'i!
..J 60
~
>
o
~
w
a:: 40
N
o
(/)
"
~
"
"
"
,," >60,.,.
"
"
~
,
20
00
I 2 3 4
STOICHIOMETRIC RATIO
5
Figure I-4. Effect of Amount and Particle
Size of Hydrated Dolomite on Sulfur
Dioxide Sorption
three times stoichiometric.
In the opinion of the authors, the incom-
plete utilization of sorbent is the result of
very rapid coating of the injected particle
with a dense sulfate coating. This opinion is
supported by the results of tests with differ-
ent particle sizes. Additional studies have
been started to study the mechanism of reac-
tion of sulfur dioxide with the sorbents.
Landesanstalt (Essen, Germanyll
Small-scale studies by Landesanstalt,
under the direction of W. Brocke, have been
carried out in a fixed-bed reactor. The ma-
terials tested, limestone and dolomite, had
the following analysis:
Limestone
Dolomite
Chemical analysis, %
Ignition loss
Cao
MgO
Fe 203
Si02
43.6
53.9
1.15
0.29
47.3
29.6
19.0
0.3
0.5
The tests were carried to the 90% breakthrough
point.
1 Brocke, W. Report for Third Limestone Sympo-
sium, Clearwater, Florida, December 4-8, 1967.
36
32
28
.
~_24
z
o
~
~20
...J
~
::)
~ 16
z
IIJ
en
It:
~ 12
8
\~'t.O - - -
c.~;:.c. --
-
;'
"
""
""
/
/
/
/
/
/
/
/
I
- LIMESTONE
- - - DOLOMITE
UNCALCINED
4
0.25
- ------UNCALCiNED--
--
-
00
Figure I-5. Effect of Gas Retention
Time on Limestone Utilization
Sorption by limestone increased with tem-
perature up to 9000 C., the highest temperature
tested. At 50% sorption, from gas containing
6 grams sulfur dioxide per normal cubic meter,
about 23% limestone utilization was obtained.
Utilization dropped rapidly at higher degrees
of removal. Calcined limestone behaved con-
siderably better; about 32% utilization oc-
curred at 50% removal (at 8000 C.). Moreover,
the calcined limestone reacted to the extent
of 24% while removing 100% of the sulfur diox-
ide, whereas only 14% of the raw limestone was
utilized at this degree of removal.
In tests of gas residence time (with dry
gas), less than 0.05 second was sufficient to
remove 90% of the sulfur dioxide but only low
loadings were possible (Fig. I-5). Limestone
gave better results than dolomite, both cal-
cined and raw. Optimum retention time was
about 0.2 second, at which about 32% loading
could be attained; little additional loading
resulted from longer retention time.
Reducing the particle size from 2-4 milli-
meters to 0.8-1.25 millimeters almost doubled
limestone utilization.
Tests without oxygen in the gas showed
that oxygen is essential to the sorption. Car-
55
-------
36
32
B
2S
~. 24
z
o
j::
~ 20
d
I-
::J
~ 16
ILl
a:J
II::
~ 12
B
.........---
"....""
".
/' C
/' .",."",----
" ,,'"
" '"
/ ,."
/ ....
/ ",'"
/ ,/"
II ,,/ PARTICLE SIZE
/" - o.S -1.25 MM.
1 --- 2 - 4 MM.
DEW POINT
A: -6. TO-9.C.
B: IS.C.
C: 50.C.
0.10 0.15 0.20
RESIDENCE TIME I SEC.
S
4
00
0.25
Figure 1-6. Effect of Water Vapor on
Sorption by Calcined Limestone
bon dioxide) however) had little effect.
Water vapor in the gas) introduced up to a
dew point of 50° C,) depressed loading ca-
pacityof calcined limestone. However) this
effect could be reduced by decreasing the
particle size (Fig. 1-6).
Resources Research Institute (Kawaguchi-
Sai tama, Japa~
Both fixed-bed and injection tests have
been made at the Resources Research Institute
in Japan) under the direction of K. Tanaka.
In the fixed-bed work) the optimum cal-
cination temperature for limestone was found
to be about 1000° C. The effect of dead-burn-
ing at higher calcination temperature could be
estimated by measurement of bulk density. Opti-
mum temperature for sulfur dioxide sorption by
calcined limestone was about 900° C.; a very
rapid decline took place below 8000 C. The
importance of small particle size was also dem-
onstrated in the fixed-bed tests.
Injection tests were made with a vertical
tubular reactor 9 meters high and 41 centime-
1 Tanaka) K. Report for Third Limestone Sympo-
sium) Clearwater) Florida) December 4-8) 1967.
56
70
60
~ 50
o
..J
~ 40
o
~
w
a:: 30
4
N
o
(/) 20
I
2
3
4
1050C
1150C
930C
840C
10
123
STOICHIOMETRIC RATIO
4
Figure 1-7. Effect of Temperature and Amount
of Sorbent on Removal of Sulfur Dioxide
with Hydrated Lime
ters in diameter. Retention time was 3 sec-
onds and flow was turbulent. Sulfur dioxide
content of the gas was 0.2 to 0.25%.
In tests with various sorbents) little
or no sorption was found after 1.5 seconds'
retention tine. With hydrated lime (median
diameter; 21 microns) the best results were
obtained at 1050° C. (Fig. 1-7); about 50% of
the sulfur dioxide was removed with a twice
stoichiometric injection.
In a test of limestone calcination rate)
more than 90% of the limestone was calcined in
0.8 second at 1000° to 1100° C. for particle
size less than 100 mesh.
In tests with limestone) at temperatures
ranging from 1010° to 1115° C,) results were
not as good as with hydrated lime. In the
best test) 20.3 and 38.1% of the sulfur diox-
ide was removed by one and two times stoichio-
metric injection respectively. Particle size
was minus 200 mesh.
Results from tests at small particle size
are shown in Figure 1-8.
-------
100
1\- I I I I I I II
t STOICHIOMETRIC
- ~ RATIO
. I
o 2
. ...
1
, ,
en
z
~ 50
u
::E
R
a::
w
I-
W
~
ex
c
w
...I
U
I-
a::
ex
a.
10
5
10
20 40 60
502 REMOVAL, %
Figure 1-8. Effect of Particle Size on
Sulfur Dioxide Sorption by Limestone
100
Steinkohlen-Elektrizitat AG (Essen, Germany)1,2
Plant-scale injection tests have been car-
ried out at the Kellerman plant of Stein-
kohlen-Elektrizitat AG (STEAG) in Germany under
the direction of K. M. Zentgraf. The boiler
used has a steam production rate of 110 tons
per hour and is of the wet-bottom type. The
coal burned in the tests contained about 1.5%
sulfur.
1 Zentgraf, K. M. Paper presented at Meeting
on Problems of Combating Flue Gas Emissons,
Munich, Germany, October 25, 1967.
2 Zentgraf, K. M. Report for Third Limestone
Symposium, Clearwater, Florida, December
4-8, 1967
Injections were made at five points, at
temperature levels of approximately 1500°,
1325°, 1150°, 1075°, and 920° C. The optimum
temperature for hydrated dolomite was found to
be 1150° C. At 2.5 times the stoichiometric
injection, about 42% sulfur dioxide removal
was obtained (including about 6% removal by
fly ash and bottom slag that occurred without
sorbent injection). Sorption fell off rapidly
as the injection temperature was increased or
decreased; 25% was obtained at 1500° C. and 21%
at 920° C.
Other materials were then injected at
1150° C.: hydrated lime, byproduct lime (from
carbide production), calcined limestone, and
raw limestone. Specific surfaces were 15.0,
10.7, 0.8, and 0.4 square meter per gram, re-
spectively (the dolomite hydrate had a specific
surface of 10.6 sq. m./g.). Particle sizes,
in terms of percent less than 20 microns, were
78, 30, 67, 48, and 73 (dolomite hydrate), re-
spectively; percentages under 100 microns were
98, 65, 98, 76, and 98.
From this it is evident that the calcined
limestone was badly dead-burned, and that the
hydrates, although high in specific surface,
contained a considerable proportion of parti-
cles larger than 20 microns. Iron content was
low (0.6% Fe203 or less) for all materials.
The order of efficiency found was hydrated
lime, dolomite hydrate, raw limestone (95%
< 90 microns), carbide lime, and calcined lime-
stone. At 1.8 stoichiometric, the hydrated
lime removed about 30% of the sulfur dioxide;
the others were considerably less efficient.
Raw limestone was then tested at 1300° and
1500° C. (Fig. 1-9). Injection at 1500° gave
much better results than at 1300° C.; in fact,
at 1500° sorption was a little better for 95%
< 90-micron limestone than for hydrated lime
at 1150° C., the optimum temperature for the
lime. A stoichiometric amount of limestone re-
moved about 20% of the sulfur dioxide.
The effect of limestone particle size was
also studied. For 70% < 90 microns, 1.4 to 1.6
stoichiometric was required to give the same
sorption as obtained with 1.0 to 1.3 stoichio-
metric of 95% < 90 microns.
Electrostatic precipitator operation was
quite good during the tests, better, it was
said, than for straight fly ash in normal op-
eration. The comparison is complicated, how-
ever, by the fact that in normal operation of
this particular boiler, fly ash was recircu-
lated to cause it to exit eventually in the
bottom slag. During sorbent injection, re-
circulation was stopped and the ash discarded
with the spent sorbent.
Zentgraf states that the Cas04-CaO-ash
mixture collected in the precipitator is suit-
able for use in road building and in block
57
-------
50
40
~
0
A
..J 30
<[
>
0
~
w
a: 20
N
0
(f)
10
PARTICLE INJECTION
LEGEND MATERiAL SIZE TEMP, @C
~~ LIMESTONE 9 5 °/0 - 90 fL 1500
~-d II Ii 1320
£ . . . .~. ...] n II i I 50
:: ','. '.' ".;
[l]]]I[1] II 10%-90fL 1500
f?/#/~ HYDRATED LIME 1150
o
0.8
2.2
2.6
1.0
1.2
1.4 1.6 1.8 2.0
STOICHiOMETRIC RATIO
2.4
Figure 1-9.
Removal of Sulfur Dioxide by Injection of
Limestone and Lime
manufacture, and that if it is discarded pollu-
tion of water supplies is not a problem.
Various methods of injection were also
studied in these tests. 1 At the 1500° C. level
the injectors were set into the wall, angled
downward and with little projection into the
boiler. At the 1160° C. injection point, wa-
ter-cooled injectors extending into the boiler
were used, with nozzles at the end of each in-
jector directing the sorbent downward against
the gas flow. Each injector served an area of
about 32 square feet. Velocities were 8.4 me-
ters per second for the main gas flow and 16
meters per second for the injection air.
At the 1070° to 1100° C. level, wall~
mounted injectors projecting only a few centi-
meters into the boiler were again used, Gas
flow was 6.3 meters per second and injection
gas flow 28 meters per second. Finally) at
1 Zentgraf) K. M. "Supplement to the Contri-
bution on S02 Measurement in Flue Gases and
on Flue Gas Desulfurization with Compounds
of Alkaline Earth Metals" (1967).
58
the 900° C. level alloy tubes containing holes
200 millimeters apart were used. The tubes ex-
tended all the way across the boiler.
Distribution at the 1500° C. point was
not very good--60% higher calcium oxide con-
centration at the boiler wall than at the cen-
ter--but could be explained on the basis of
boiler configuration rather than injector ar-
rangement. The similar wall-mounted injectors
at 1070° to 1100° C. did somewhat better--a
ratio of 1:1.35 in extreme values and the high
values were scattered over the boiler width.
Best results were obtained for the counter-
current injection at the 1160° C. level; ex-
treme values were said to be not so pronounced
but no numerical values were given.
It should be noted that the distribution
was not measured at the particular injection
point but at a somewhat remote point almost
at the end of the convection pass. No data
are available on the time required after in-
jection for the measured distribution to be-
come established.
Other figures given by Zentgraf have a
-------
bearing on the effect of retention time. In
dust sampling, the measured sorption was af-
fected by retention time in the sampling probej
an increase from 0.05 second to 0.25 second in-
creased the indicated sulfur dioxide removal
from 7.5 up to 12.5%.
Technical University of Stuttgart, Institute
of Process Technology (Stuttgart, Germa~
Plant-scale tests of limestone injection
have been carried out at a BASF (Badische
Anilin- und Soda-Fabrik AG) power plant in
Ludwigshafen, Germany, by the Institute of
Process Technology (Technical University of
Stuttgart), under the direction of K. Gold-
schmidt. The boiler was of the dry-bottom,
two-pass, natural circulation type, with a
maximum steam output of 54 tons per hour.
tention time in the boiler was relatively
4 to 5 seconds from the 1000° C. point to
air heater.
In the first test series, hydra.ted lime
was fed with the coal. The lime dead-burned
severely and therefore sorbed sulfur dioxide
poorly. Moreover; the lime reduced the fly
ash fusion temperature and caused severe slag-
ging in the convection section.
Actual removal in these tests was 13 to
20% at stoichiometric and 22 to 36% at 2.8
times stoichiometric. This includes about 5%
removal by the fly ash when no lime was added.
Results from injection at 950° to 1000° C.
were much better (Fig. 1-10). Over 30% removal
was accomplished with hydrated lime and there
was little increase in boiler slagging. The
hydrated lime was considerably more effective
than the hydrated dolomite. At 2.8 stoichio-
metric and 1000° C., the lime removed over 50%
of the sulfur dioxide.
Sorption was further improved when oil
was fired rather than coal (Fig. 1-11). Sorp-
tion by hydrated lime increased to 40% at stoi-
chiometric and 77% at 2.8 times stoichiometric.
The lime was again far superior to the dolomite.
The better results with oil firing were
attributed to the increased retention time re-
sulting from sorbent depositing on surfaces in
the boiler. It was postulated that this is
beneficial in oil firing but that the molten
fly ash that also deposits in coal firing cov-
ered the deposited sorbent and prevented sul-
fur dioxide sorption. Evidence for this was
the fact that in oil firing the degree of sorp-
tion increased gradually over the first hour
of injection, indicating that sorbent building
up on surfaces was assisting in sulfur dioxide
removal. This effect was less pronounced at
Re-
long,
the
1 Goldschmidt, K. Report for Third Limestone
Symposium, Clearwater, Florida, December
4-8, 1967.
high stoichiometric ratios, which would be ex-
pected because of the higher ratio of sorbent
carried in the gas to sorbent deposited on the
walls. In contrast, sorption increased rapid-
ly to the maximum in a minute or so when the
boiler was fired with coal.
The poor results with dolomite hydrate
were explained on the basis of (1) higher de-
composition pressure of magnesium sulfate at
the optimum calcium oxide reaction temperature
(plus, presumably, slow reaction rate at the
lower temperature optimum for magnesium oxide
reaction with sulfur dioxide), (2) lower sur-
face area of dolomite hydrate (no values were
given), and (3) high impurity content of the
dolomite used (no analysis given).
It was noted that in most of the runs the
sorbent reacted with carbon dioxide as well as
sulfur dioxide.
Efforts were made to inject at a third
point, in the "throat" of the boiler just be-
fore the convection pass, where the tempera-
ture was about 850° C. The poor results were
explained on the basis that the high gas ve-
locity at this point prevented uniform disper-
sion of the sorbent. No data were given on
uniformity of distribution for any of the in-
jection points.
Central Research Institute of Electric Power
Industry (Tokyo, Japan)2
Both pilot plant and full-scale limestone
injection studies have been carried out by the
Central Research Institute of Electric Power
Industry in Japan, under the direction of Y.
Ishihara. The program began in 1964 and is
continuing at present.
In the pilot plant studies, the optimum
temperature for limestone injection was found
to be 900° to 1200° C. (which is not consis-
tent with Zentgraf's findings). Most of the
sorption occurred during particle travel of 6
meters, during which the temperature dropped
from 1100° to 800° C. Good removal was ob-
tained--32 to 40% at stoichiometric. However,
about 15% removal was noted in tests without
limestone, attributed to sorption on wall sur-
face s .
Removal decreased inversely with the one-
fourth power of particle size. Limestones
containing over 1% ferric oxide were about 50%
more effective than those containing less than
0.5%. Data on particle size and iron content
are shown in Figure 1-12. In all these pilot
plant tests the boiler was oil fired.
In the plant tests, in an oil-fired boil-
er, the sorbent was injected near the top of
2 Ishihara, Y. Report for Third Limestone
Symposium, Clearwater, Florida, December
4-8, 1967.
59
-------
60
~ 50
o
...
....I
~ 40
o
~
~ 30
a:::
G: 20
....I
::>
(f) 10
o
o
80
70
~ 60
o
..
...J
~ 50
o
~
w 40
a:
a:::
~ 30
....I
::>
(f) 20
10
°0
60
o LIME TEST
. DOLOMITE TEST
0.5
1.0
1.5 2.0 2.5 3.0
STOICHIOMETRIC RATIO
3.5
4.0
4.5
Figure 1-10. Sorption of Sulfur Dioxide by Hydrated Lime and
Hydrated Dolomite Injected into a Coal-Fired Boiler
o
o LIME TEST
. DOLOMITE TEST
0.5
1.0
1.5 2.0 2.5 3.0
STOICHIOMETRIC RATIO
3.5
4.0
4.5
Figure I-II. Sorption of Sulfur Dioxide by Hydrated Lime and
Hydrated Dolomite Injected into an Oil-Fired Boiler
-------
I I I
° Fe203' 0 - 0.5%
II Fe203. 0.5 - 0.1%
0 Fe 2 03. ABOVE 1.0 %
~ ~ 0
0
~ - """"'
........ ...... II II ° [~
...... ...
II -.............. II
II
r--- II
IlO I..............
-
100
90
80
70
60
50
~ 40
~
u
~ 30
0::
o
o
U
LL 20
o
~
z
UJ
U
0::
UJ
11..
10
9
8
1
6
5
1
3
4 5 6 7 8 9 10
PARTICLE SIZE OF LIMESTONE
50 60
2
20
30
40
Figure 1-12.
Relation Between Rate of Reacted Lime and
Particle Size of Limestone
the boiler at a gas temperature of 11000 to
12000 C. Results for three particle sizes of
limestone and for hydrated lime are shown in
Figure 1-13. Hydrated lime appeared to be
superior, and in the tests at twice stoichio-
metric the 10-micron particle size was a little
more effective than the larger sizes.
Dust deposits taken from the wall and tube
surfaces varied in sulfate content depending
on the point of collection. At points where
the gas temperature is normally about 9500 C.,
from 80 to over 95% of the calcium oxide had
been converted to sulfate. As the temperature
decreased the calcium sulfate content was al-
so lowered. Below 5000 C. the content was uni-
form, at about 40%; presumably these solids
were already reacted when they deposited.
There was some difficulty from the deposits
on the tubes interfering with heat transfer.
Wisconsin Electric Power Company (Milwaukee,
Wisconsin) 1
During 1964, Wisconsin Electric Power
Company initiated studies to reduce the sulfur
dioxide emission from coal-fired power gener-
ating stations. As part of these studies,
tests of direct injection of limestone into
a boiler were made at the Port Washington
(Wisconsin) plant.
The plant consists of five 80-megawatt
units. The boilers, built by Combustion Engi-
neering, Inc., are equipped with 20 vertically
fired burners. Limestone was added to the
No.1 unit; No.2 was fired under similar op-
erating conditions but without limestone.
The limestone used was a dolomitic mate-
1 Pollock, W. A., Tomany, J. P., and Frieling,
G. Air Engineering 2. (9), 24-28 (1967).
61
-------
60
. NON-I NJECTED .
0 LIMESTONEUOp)
!;. LIMESTONE (141')
0 LI MESTONE (17 1')
. HYDRATED LIME 0
o
0
0
~
.
...J
«
>
o
::E
~ 40
w
c
X
Q
c
II::
;j
~ 20
;:)
U)
o
o
123
INJECTED WEIGHT (RATIO TO STOICHIOMETRI C)
Figure 1-13. Relation Between Sulfur Dioxide
Removal and Theoretical Equivalent of
Limestone for Sulfur in the Fuel Oil
rial obtained locally. The coal, from a West
Virginia mine, contained about 2.8% sulfur and
8% ash. The limestone and coal were propor-
tioned continuously by weigh feeders onto a
conveyor belt from the storage yard.
Three tests were made with a 90% coal-10%
limestone blend and one test with a 95% coal-
5% limestone blend. These proportions were
equivalent to about 130% and 65% of the re-
quired calcium oxide to react with the sulfur
dioxide in the flue gas.
The conveyor belt from the storage yard
discharged the limestone-coal mixture into a
primary crusher, which was not required for
size reduction but served as a blender for the
limestone and coal. The crusher discharged
into coal pulverizers which ground the mix to
about 65% minus 200 mesh. Operation of the
pulverizers was not affected by the presence
of the limestone. The pulverized mix was
stored in a bunker and fed from the bunker by
a star-type feeder.
The limestone-coal mixture was fired in
the No.1 boiler for several days to ensure
steady-state operation. Samples of the flue
gases from No.1 and No.2 units were taken
simultaneously. Analyses of these samples in-
dicated that during the periods the No. 1 unit
was fired with mix containing 10% limestone,
the sulfur dioxide emission rate was about 50%
less than the emission rate from the No.2
unit (Table I-I). When the mix with 5% lime-
stone was fired, no significant difference was
detected in the sulfur dioxide emission from
62
the two units. However, the ash picked up
sulfur trioxide in these tests, indicating
that sorption was taking place.
The use of limestone in the No.1 furnace
resulted in the formation of a deposit on the
superheater tubes which could not be removed
by the soot blowers. After the tests were com-
pleted, it was necessary to manually scrape
the deposit from the tubes. The material de-
posited was quite fragile when cold but ex-
tremely sticky and tenacious at elevated tem-
peratures. Data on effect of limestone on ash
fusion temperature are given in Figure 1-14.
4
Combustion Engineering and Detroit Edison
(Detroit, Michigan)l,2
After a series of pilot plant tests, Com-
bustion Engineering, in cooperation with De-
troit Edison, conducted a large-scale test of
dry injection of dolomite into a boiler follow-
ed by wet scrubbing. Into one side of a twin
325-megawatt furnace there was injected the
stoichiometric amount of dolomite to remove
the sulfur dioxide formed by combustion of 3
to 4% sulfur coal. The dolomite was injected
through the top row of burners (four rows in
all), after being ground in one of the plant
coal pulverizers.
No interference with normal steam output
occurred during 30 days of continuous addition
of dolomite. Most of the injected dust (over
90%) left the boiler with fly ash rather than
bottom ash.
The degree of sulfur dioxide sorption
varied with tilt of the burners, which were of
the type that can be tilted to control boiler
temperatures. In all tests the dolomite in-
jector "burners" were tilted upward 15° above
horizontal. Results at various coal burner
tilts were as follows:
Divergence of
coal burner tilt,
degrees relative
to horizontal
Maximum (-30)
Medium (-15)
Minimum (+15)
Minimum (+15)
S02 content
of gas, p.p.m.
Without With
dolomite dolomite
S02 re-
moval, %
28.8
21.5
21.2
14.8
2460
2556
2460
2591
1755
2015
1939
2242
The dust loading in the stack gas increas.,
ed about 65% as a result of the dolomite in-
jection. Mechanical collectors ahead of the
l Plumley, A. L., Whiddon, O. D., Shutko, F.
W., and Jonakin, J. "Removal of S02 and Dust
from Stack Gases." Paper presented at Amer-
ican Power Conference, Chicago, Illinois,
April 25-27, 1967.
2 Research-Cottrell, Inc. Private communica-
tion (1967).
-------
Table I-I.
Wisconsin Electric
Sample
No.
--
Boiler 2 (no limestone)
802 in r;as, 1, S03 in ash, ~,
Run 1
1
2
3
4
5
1.54
1.84
1.85
2.05
1.70
0.117
0.104
0.113
0.124
0.130
Run 2 (5% limestone in mix)
1
2
3
0.136
0.112
0.122
1.38
1.21
Run 3
1
2
0.140
0.125
4.31
Run 4
1
2
3
4.22
4.06
4.41
0.135
0.104
0.102
a
Tests on Sulfur Dioxide Sorption
a S02
Boiler 1 (with limestone) removal,
S02 in gas, "/ SO,,\ in ash, ~, %
0.058
0.053
0.055
0.081
0.068
5.06
5.61
5.40
5.57
5.7)+
50
4)
51
35
40
0.125
0.108
0.119
5.43
5.47
8
4
3
0.083
0.089
8.21
41
29
0.072
0.059
0.055
6.32
7.46
6.52
47
43
46
10~ limestone in coal-limestone mix unless otherwise noted.
precipitator operated with no loss of efficien-
cyj particle size distribution in the dust was
about the same as normal fly ash. However,
performance of the electrostatic precipitator
was seriously affected by the change in dust
composition and loading. The electrical char-
acteristics changed almost immediately from
loaded to sparking conditions ana collection
efficiency dropped from about 80% (normal) to
55%. Operating temperature was about 2700 F.
Overall collection efficiency, mechanical and
electrostatic, dropped from 96% to 90%. A
summary of the composite data on precipitator
operation is shown in Table I-II.
Both high-temperature and low-temperature
corrosion measurements were made during the
injection period. Data from test samples in
the air heater region were as follows:
Average
weight
loss, g.
S03 in
gas,
p.p.m.
o
17
With dolomite
Without dolomite
0.0078
0.0456
pH of
deposits
10.8
3.2
Specimens of various superheater materials
were suspended in areas where liquid phase
corrosion is known to occur. The results are
shown in Figure I-15. Corrosion with dolo-
mite injection was little above that expect-
ed from ordinary oxidation at the temperatures
involved.
No unusual slagging tendencies were ob-
served, even though a high-ash (14%) coal was
used during part of the injection period. Lab-
oratory tests were run on ash from both this
coal and a more normal one (10-13% ash) to de-
termine effect of dolomite addition on ash fu-
sion temperature. For the normal coal, the
dolomite decreased ash-softening and fluid tem-
perature up to about a 45:55 dolomite-ash mix-
ture. At 50:50, however, a ratio that would
be reached even at stoichiometric ratio for
high-sulfur coals, the temperatures were high-
er than without dolomite. For the low-grade
coal the same result was noted for the fluid
temperaturej softening and deformation tem-
peratures, however, rose even at small dolo-
mite additions.
63
-------
2500
LL
o
...2300
w
a:::
::>
t-
«
a:::
w
Q.
~ 2100
t-
F'L.UIDIZIIVG
..
.
.
2000
1900
o
2
3
4 5 6 7
ASH CALCIUM CONTENT. 0/0
8
20
Figure 1-14.
Effect of Limestone Injection on Ash-Fusion Temperature
Table I-II. Effect of D~lomite Injection on Dust-
Collection Efficiency in Combustion
Engineering - Detroit Edison Tests
Dolomite injected, tons/hr. 6
802 in stack gas, % 0.195
803 in stack gas, % Nil
Dust resistivity, ohm-em. 1022
Gas volume, c.r.m. 246,000
Gas temperature at precipitator 287
Moisture in gas, % 6.2
Dust concentration, grains/s.c.f.
Into mechanical collector
From precipitator
Precipitator inoperative
Precipitator operating
Dust collection efficiency, %
Overall
Precipitator inoperative
Precipitator operating
Precipitator only
o
0.255
0.0017
108
284,000
270
7.2
6.1
3.7
0.74
0.16
1.32
0.60
78
90
55
80
96
79
64
-------
600
500
C)
~.. 400
(J)
(J)
o
..J 300
t-
:I:
C)
~ 200
4 WEEKS EXPOSURE
WITHOUT DOLOMITE
.
.
III
.
~A
.
A
IP
II
A
100
5 WEEKS EXPOSURE
WITH DOLOMITE
A
eo:: 347 5.5.
II 0 ::::321 5.5.
A A =316 5.5.
o
900
1000
1100
TEMPERATURE, of.
1200
Figure 1-15.
Corrosion Rate of Test Samples Exposed
in Superheater Region
1.26
1.05
..J
.84 :?:
..
z
o
.63~
0::
t-
W
z
.42~
.21
o
1300
-------
APPENDIX II
LIMESTONE AVAILABILITY AND TECHNOLOGY
If limestone injection should become a
widely used method for sulfur oxide control in
power plants, a large quantity of limestone
would be needed. In fact, if half the 15 mil-
lion tons of sulfur emission expected from
power plants in 1980 were removed from the
gases by limestone, from 24 to 48 million tons
--the range from 100 to 200% stoichiometric--
would be required. Moreover, the limestone
should be economically minable, well located
in relation to power plants, and of adequate
quality.
Therefore, the extent, accessibility,
geographical location, and chemical-physical
properties of limestone reserves (both active
and undeveloped) are all important factors in
evaluating the competitive status of the in-
jection method. Available information on
these points is summarized in this appendix.
And since calcination or calcination-hydration
before injection may be advantageous, modern
practice for these operations is summarized.
Much of the information presented has
been obtained by discussions with people in
the limestone industry. In addition, the
assistance of R. S. Boynton, Director of the
National Lime Association, through both direct
discussions and his reference book on lime-
stone,l is acknowledged. Recent published
literature on the subject has also been sur-
veyed.
Location and Nature of Deposits
Limestone deposits are prevalent in prac-
tically all the states of the United States
and in most of the countries of the world.
Economically minable limestone does not occur
so frequently, however; because many of the
deposits are too small, too impure, or too
deeply embedded in the earth. Although there
are maps and other information on total depos-
its, there apparently is no good information
as to which of the undeveloped deposits are
economically minable. The locations of active
deposits are shown in the accompanying figures
--high-calcium limestone in Figure 11-1 and
dolomitic and high-magnesium limestones in
Figure 11-2.
Although there are no good figures on the
tonnage in economically minable reserves, it
is obviously quite large. Reserves of chemi-
cal grade (95% or more total carbonates), how-
ever, have been estimated to be only about 2%
1 ~~~:t~~~ ~~m~~to~::m~~~:~s~~~n~:c(~~~61~ of
66
of the total reserves. As will be discussed
later, it may be that chemical grade, or some-
thing close to it, will be required to get
adequate reactivity for sulfur dioxide sorp-
tion. The large annual consumption of chemi-
cal grade is draining the reserves rapidly.
Beneficiation, a relatively expensive proce-
dure, is already being practiced in some
plants.
In discussions with limestone producers
during the course of this survey, several sit-
uations were found in which the adequacy of
reserves was in question. These were all re-
lated to chemical-grade limestone; reserves of
lower grade suitable for cement manufacture or
construction purposes appear ample. In some
cases, the local reserve of high-grade material
was near exhaustion and in others the need for
more expensive mining methods (shaft mining or
removal of heavy overburden) was imminent.
The active deposits (Figs. 11-1 and 11-2)
are located mainly in the central and north-
eastern sections of the country, presumably
because these are the areas of principal de-
mand. Many of the blank spots in the active
deposit maps have limestone deposits but they
are not mined because of limited need. The
developed deposits coincide fairly well with
location of power plants (Fig. 11-3), again
probably because these industrialized regions
require large amounts of both limestone and
electric power. The main advantage of the co-
inciding concentrations is that limestone
would be readily available, without any period
of quarry development, if power companies
adopted limestone injection for sulfur dioxide
control.
Active dolomite deposits are more widely
scattered than the high-calcium type. There
are supplies of high-calcium limestone, either
developed or undeveloped, near practically all
the power plants.
The main problem in regard to limestone
supply appears to be in the coastal areas,
where natural deposits are relatively rare.
Shellfish operations produce supplies of cal-
cium carbonate in such areas but these may not
be adequate. A possible source for coast lo-
cations is a large deposit of aragonite lo-
cated in Bermuda, now being mined and shipped
to coastal and river points. Ocean shipping
costs are relatively low as compared with rail.
In some parts of the countrY--Tennessee-
Mississippi-Alabama, Texas-Arkansas, and
Kansas-Nebraska-South Dakota--there are depos-
its of chalk, a soft limestone of very small
-------
--~- --
'"
"'.
",.
.",
"..
"
,
- -;-- - ~ - -- ----
:-- - -- --.- - - -----
,
-- - - - ---
..
---:-----------
-----,--J_----
:--------
.:
--J -_:-- - - - - ; - --
).
..'
------'
" L F
o F
M E X J C 0
,..
----
".
.,;
Figure 11-1. Location of Active High-Calcium Limestone Deposits
LFrom Chemistry and Technology of Lime and Limestone, R. S.
in the United States
Boynton
(196617
0\
-l
-------
II. OAK.
S. OAK.
.
NEB.
KAN.
.
.
Figure 11-2.
Deposits in
Location of Dolomitic and High-Magnesium Limestone
the United states, Most of Which are in Operation
LFrom Chemistry and Technology of Lime
and Limestone, R. S. Boynton (196617
crystal size reported to be easier to grind
than ordinary limestone. These properties
might make chalk a superior sorbent for sulfur
dioxide. Unfortunately, it is reported that
must of the chalk deposits are relatively low
in grade, some containing 20 to 30% silica
plus alumina. One of the deposits, near
Demopolis, Alabama, was visited during the
course of this survey. The chalk has little
or no overburden, is about 1100 feet deep, and
is quarried by scrapers working against a ver-
tical face. Moisture content is relatively
high (about 16%) so the material is dried be-
fore fine grinding, which is carried out in a
ball mill. Because of the soft nature of the
chalk, mill capacity is high and power re-
quirement low. The low mining cost should
make the price of the material relatively low;
the product is all captive at present (cement
production) but it was indicated that sale for
power plant use could be arranged, perhaps at
a price as low as $0.75 per ton.
Limestone Composition and Quality
Impurities appearing in quarried lime-
stone may be either an integral part of the
deposit or material from crevices or other
strata excavated along with the limestone.
Silica and alumina, the major impurities, usu-
68
ally come from the latter source, and iron ox-
ide, the third most common, may be either ho-
mogeneously incorporated or obtained inciden-
tally as particles of pyrite or limonite. Or-
ganic matter; usually under 1%, is often found.
Typical compositions of limestone types
are given in Table II-I.
The composition requirements for sulfur
dioxide sorption are not clear. None of the
investigators so far has found a clearcut re-
lation between sorption efficiency and type or
quantity of impurity. The nearest to this
probably is NCAPC-TVA work in which limestones
containing relatively high amounts of silica
and alumina had relatively low porosity (and
low sorption efficiency) after calcination,
presumably because of slag formation that
blocked pores. This is in line with industrial
experience in lime burning, as will be dis-
cussed later. No quantitative data on the ef-
fect are yet available.
Iron may be a beneficial impurity since
investigators such as Ishihara have found it
helpful, presumably because it catalyzes sul-
fur compound oxidation. Ishihara obtained
considerably better sorption with limestones
containing over 1.5% ferric oxide than with
those containing w1der 0.5%. No data have
been found on the incidence and extent of high-
-------
o
Figure 11-3.
.
Approximate Location of Major Thermal Power Plants
iron deposits. In Table II-I, only one of the
samples (No.2) can be called high iron and it
is so high in silica that the calcium oxide-
magnesium oxide content is substantially dilu-
ted.
The effect of composition of various
limestones on sulfur dioxide sorption is being
studied by several organizations. If their
findings indicate that low silica and alumina
contents are necessary, the cost and availabil-
ity of limestone for sulfur dioxide sorption
will be adversely affected. Like chemical-
grade limestone, only certain deposits will be
suitable and in many situations it will be
necessary to quarry selectively, with conse-
quent increase in cost.
Commercial Production and Shipping
Limestone is a major commodity in the
United States, ranking second after sand and
gravel. The total tonnage in 1965 was 554
million tons (sales value of $747,000,000),
over half of which was produced as crushed and
broken stone used as aggregate in concrete and
road beds. The number and size of commercial
operations can be approximated from Table
II-II; the values given are for crushed stone,
of which about 70% is limestone.
The large current tonnage is favorable
for sulfur dioxide removal by limestone be-
cause the large amounts required for injection
would not have a major and upsetting impact on
the supply situation, assuming that the aver-
age grade is suitable. The 48 million tons
required to remove half the 1980 emission (as-
suming 200% stoichiometric) would amount to
only a little over 10% of the total, even on
the basis of 1965 consumption.
Of the 50 states, 47 produce limestone.
The leading producers in 1965 are shown in the
following tabulation.
69
-------
Table II-I . Representative Compositions of U. S. Limestones a
Typesb
1 2 -L 4 ----L- 6 -1- 8
CaO 54.54 38.90 41. 84 31.20 29.45 45.65 55.28 52.48
MgO 0.59 2.72 1. 94 20.45 21.12 7.07 0.46 0.59
CO2 42.90 33.10 32.94 47.87 46.15 43.60 43.73 41. 85
Si02 0.70 19.82 13.44 0.11 0.14 2.55 0.42 2.38
A1203 0.68 5.40 4.55 0.30 0.04 0.23 0.13 1. 57
Fe203c 0.08 1.60 0.56 0.19 0.10 0.20 0.05 0.56
SO 3d 0.31 0.33 0.33 0.01
P20s 0.22 0.05 0.04
Na20 0.16 0.31 0.06 0.01 0.01
K20 0.72 0.01 0.03
H20 1. 55 0.16 0.23 n.d.
Other 0.29 0.01 0.06 0.08 0.20
a Boynton, R. S., Chemistry and Technolo~y of Lime and Limestone,
Interscience (1966).
b 1 = Indiana high-calcium limestone.
2 = Lehigh Valley, Pennsylvania "cement rock."
3 = Pennsylvania "cement rock."
4 = Illinois Niagaran dolomitic limestone.
5 = Northwestern Ohio Niagaran dolomitic limestone.
6 = New York magnesian limestone.
7 = Virginia high-calcium limestone.
8 = Kansas cretaceous high-calcium limestone (chalk).
c Some Fe reported as FeO.
d Some S reported as elemental S.
Table II-II. Number and Production of Crushed Stone Plants
C umulat i ve
Number total,
Annual production of Thousand Percent thousand
(short tons) plants short tons of total short tons
Less than 25,000 889 7,462 1.3 7,462
25,000 to 50,000 292 10,805 1.8 18,267
50,000 to 75,000 241 14,828 2.5 33,095
75,000 to 100,000 193 16,899 2.9 49,994
100,000 to 200,000 499 71,421 12.1 121,415
200,000 to 300,000 248 60,554 10.3 181,969
300,000 to 400,000 187 65,654 11.1 247,623
400,000 to 500,000 137 61,046 10.4 308,669
500,000 to 600,000 94 50, 556 8.6 359,225
600,000 to 700,000 50 32,325 5.5 391,550
700,000 to 800,000 36 26,900 4.6 418,450
800,000 to 900,000 21 17,627 3.0 436,077
900,000 tons and over 92 152,302 25.9 588,379
Total 2,979 588,379 100.0 588,379
70
-------
Millions of tons
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Pennsylvania
Illinoi s
Ohio
Missouri
Michigan
Florida
Tennessee
Texas
New York
Kentucky
Iowa
Indiana
Virginia
47.9
47.0
41.5
34.9
34.6
34.3
28.8
27.9
27.5
26.0
25.9
24.0
21.9
Limestone is produced in crushed, broken,
and ground form. Purity, particle size, and
particle-size gradation vary according to end
use. The following tabulation gives 1965 data
on end use and price.
Concrete and road
Cement manufacture
Flux
Agricultural
Lime manufacture
Rip rap
Ballast
Short tons,
1,000' s
337,770
92,110
32,530
28,289
26,460
13,874
6,150
Price/short
ton, $, Lo.b.
1.30
1.04
1.44
1.72
1.67
1.06
1.72
Limestone for flux, agriculture, and lime
manufacture is usually high-purity material
(95% or more combined carbonates). Purity is
not important in construction-grade stone; the
price variation shown reflects difference in
particle size. Cement stone must be low in
magnesium but can contain silica, alumina, and
iron.
Shipping distances for limestone are rel-
atively short, usually a maximum of 75 to 100
miles. There is some equalization of freight
but the low f.o.b. prices cannot stand much
freight absorption. Because of the short
hauls, most limestone moves by truck as shown
in the following tabulation.
1963
Million Percent
short tons of total
Method of
transportation
Commercial
Truck
Rail
Water
Unspecified
Total
4n
83
51
67
618
Government and
contract truck
Total
67
685
61
12
7
10
90
10
100
Many commercial operations include both
quarrying of raw stone and calcining to make
lime or hydrated lime. Although such opera-
tors sell raw limestone, they are probably net
the best sources of limestone for sulfur diox-
ide control because their operations are rela-
tively small and the limestone they quarry
must be of premium quality for lime production.
The crushed stone producers, many of whom sell
over 900,000 tons per year; are more likely
sources of low-cost material. In the survey
conducted as part of this study, one such pro-
ducer quoted a price of only $0.71 per ton
f.o.b. for 100,000 or more tons per year. The
output of this particular producer is over 5
million long tons per year, in a plant capable
of 2000 long tons per hour. Modern equipment
and large volume make the low selling price
possible.
One possibility for reducing cost is to
use limestone that is a byproduct from other
operations. There are three main types: (1)
limestone overburden removed in ore mining,
(2) low-grade limestone removed in selective
mining of high grade, and (3) fine material
separated in producing limestone sized to the
customer's requirements. Sometimes very large
quantities of such byproduct are available at
low cost. For example, a large ore mining op-
eration in the South has produced over the
years a 3 million-ton pile of limestone that
is available at about $0.65 per ton, about
half the usual price for volume sales in the
area. In another instance, a company mining
high-calcium limestone removes dolomitic lime-
stone as a waste product, at a rate of several
hundred thousand tons a year. This also is
available at relatively low price. Still an-
other case is a rip rap producer who would
sell the fines from this operation for $0.65
to $0.75 per ton (85-88% CaCOs); $0.50 per ton
was said to be the quarry operation cost.
For the large quantities needed in lime-
stone injection, a special quarrying operation
might well be appropriate. For example, the
600,000 tons or more per year needed for a
lOOO-megawatt power plant (3.5% S; twice stoi-
chiometric) would exceed the annual capacity
of 95% of the crushed stone plants in the
United States. Hence it would appear logical
to set up a quarrying operation at the closest
suitable deposit to supply a given power plant.
With modern large quarrying and crushing equip-
ment, costs should be relatively low.
Grinding
Some of the test data have indicated that
fine grinding of limestone, to a fineness on
the order of 99% minus 325 mesh (approximate
average of 8-10 microns), will be justified
economically as a means of increasing sorbent
71
-------
utilization. Limestone of this fineness is
equivalent to whiting, which is finely ground
limestone used as filler in products such as
paint, paper, and rubber.
Whiting tonnage was only 818,000 tons in
1965. If a 48 million-ton consumption for
limestone injection developed as mentioned
earlier; fine grinding would become a major
operation.
Since the cost of the fine grinding is a
major consideration, the whiting industry was
surveyed during the course of this study to
determine grinding practice. A wide variety
of practice was found. Ring-roll, ball, and
hammer mills are used, most of them air swept.
Contrary to published reports, the manufactur-
er of one of the fluid energy type of mills
stated that this type is not suitable for lime-
stone. Only one wet-grinding installation
(ball mill) was found.
Typical grinding situations found were as
follows:
1. Air-swept ring-roll mill
99% minus 325 mesh
Price: $11 to $12 per ton f.o.b.
2. Air-swept ring-roll mill
99.5% minus 325 mesh
Price: $11 per ton f.o.b.
3. Wet grinding in ball mill
100% minus 325 mesh, 85% minus 7 mi-
crons, 39% minus 2 microns
Price: $37 per ton f.o.b.
4. Dry grinding in ball mill
Average of 2 microns
Power consumption and capacity reported
for these operations agreed fairly well with
the published figures used in the grinding
cost estimate (Appendix IV). Sales price was
similar, in most cases, to the $12 per ton
national average for whiting in 1965; no data
could be obtained on actual production cost.
The prices quoted are far higher than would be
indicated by the cost of grinding estimated in
Appendix IV. However; low volume and the spe-
cialty nature of the market probably account
for this.
It is concluded that either ring-roll or
ball mills are suitable for fine grinding of
limestone. As pointed out earlier, ball mills
should be more economical for large-scale
grinding. Wet grinding, which appears not to
be used very much in the whiting industry
(presumably because of the product drying
cost), may be preferable for power plants if
slurry injection is found to be feasible.
Practice in grinding to intermediate par-
ticle size, such as the 70% minus 200 mesh as-
sumed in the base case for the estimates, was
also surveyed. This is a prevalent size in
grinding of phosphate rock, which is similar
to limestone in grindability. Air-swept ring-
72
roll mills with internal classifiers are used
almost exclusively in the phosphate industry
for intermediate-size plants. For modern,
very large plants, air-swept ball mills are
the usual choice.
The cement industry has broad experience
in grinding of limestone to intermediate size.
No instance of ring-roll mill use was found,
presumably because all the operations checked
produced a large tonnage, for which ball mills
should be more economical. The typical grind-
ing installation in the cement industry is a
ball mill with gravity discharge and a bucket
elevator to an external classifier, with re-
cycle of oversize from the classifier back to
the mill.
Calcination
Limestone calcination practice was sur-
veyed to get information for preparing esti-
mates on cost of calcining before injection.
In addition, information both on practice and
theory of calcination is needed for optimizing
injection of raw limestone.
As has been discussed earlier, proper
calcination of limestone requires careful se-
lection and control of temperature and reten-
tion time. Most of the published data--and
the current practice also--are related to cal-
cination of fairly large lumps. This is ap-
plicable if the limestone is to be calcined
before injection, but the time-temperature
situation is quite different when the calcina-
tion occurs in the boiler; the temperature is
higher; the retention time is of a different
order of magnitude, and the particle size is
smaller.
Little information appears to be available
on calcination of particles as small as those
contemplated for limestone injection, particu-
larly if 99 to 100% minus 325 mesh is the size
selected. The work of Butt and Stolovitskaya
in Russia seems to be the most pertinent. Butt,
et al.l tested finely divided limestone 1three
fractions: 1-3 mm., 0.5 (30 mesh)-l mm., and
-30 mesg7 in a fluidized calcining system.
Calcination required only 0.05 to 0.038 second,
depending on the calcination temperature. Par-
ticle size was decreased during calcination,
presumably because of thermal shock.
Other investigators have noted that some
limestones disintegrate when heated. Appar-
ently limestones with large crystal size are
more subject to this than those with a fine-
grained structure. No information was obtained
in the survey on location of deposits that
l Butt, Yu. M., Kolbasov, V. M., and Lagoida,
A. V. Eksperiment v Tekhn. Mineralog. i
Petrogr., po Materialam Soveshch., 7th, Lvov
1964, 265-271 (Pub. 1966) (Russ).
-------
contain the large crystal type. Boynton
states that limestones can be grouped into
four categories.
1. Those that fracture and decrepitate
readily during preheating and low calcina-
tion temperatures.
2. Those that will yield a porous, reac-
tive lime under most calcination conditions
and which are difficult to overburn.
3. Those that will yield a dense, unreac-
tive lime of low porosity even under the
mildest calcining conditions.
4. Those that yield a porous, reactive
lime under mild temperature conditions and a
denser, less porous lime under harder burn-
ing conditions.
It would appear desirable to identify deposits
that contain the first two types because they
should be superior for limestone injection.
Stolovitskayal also calcined in a fluid-
ized bed, at a fixed retention time (0.25 sec-
ond) and at temperatures of 1100° to 1400° C.
Recrystallization of calcium oxide increased
with temperature; crystal size at 1100° C. was
too low to measure but at 1400° the crystals
were as large as 2 to 4 microns. Calcination
was complete in the 0.25 second at 1200° C.
and higher. Reactivity declined with increase
in temperature; the hydration period was 5
minutes for 1200° C. material and 16 minutes
for 1400° C.
In other work, with particles of larger
size in fixed beds, Hedin2 found that dead-
burning is not a straight line function of
temperature; there was actually less burning
at 1150° than at 1050° C. in 2-hour tests, but
as temperature was increased above 1150° C.
burning increased rapidly to above the 1050°
level.
Tagawa, et al.3 (Nippon Carbide Ind. Co.,
Ltd., Japan) found that reactivity of calcine
is closely related to bulk density.
Other investigators hold that rate of
heating has a greater effect on reactivity
than temperature or retention time, and that
slow rather than shock heating is preferable.
Shock calcination is said to densify the re-
sulting calcine and reduce porosity. If so,
the effect is unfortunate for little can be
done in regard to heating rate of injected
limestone.
There is general consensus that tempera-
ture level has a much greater effect on cal-
cine reactivity than does retention time.
Temperature also has a greater effect on cal-
cination rate, and limestones vary widely in
optimum temperature for getting the best com-
bination of calcination rate and reactivity.
Magnesium carbonate is less subject to dead-
burning than calcium carbonate, as growth of
magnesium oxide crystals is much slower.
Hence dolomite usually is more porous than
high-calcium limestone when both are calcined
at optimum temperature.
The effect of impurities on calcine reac-
tivity is not well worked out. Murray4
studied 43 limestones and found no correlation
of composition with shrinkage (an inverse mea-
sure of reactivity), except that sodium de-
creased shrinkage. Boynton points out that
low calcining temperature is necessary to get
even moderate activity with impure limestones,
and Murray obtained better reactivity with Ice-
land Spar; a very pure limestone, than with
less pure types. This is presumably due to
the effect of silica and alumina, which are
not highly reactive at low temperature but may
slag calcium at high temperatures and as a re-
sult block pores.
The beneficial effect of sodium in the
limestone is such that several have proposed
or tested addition of a material such as sodi-
um chloride or sodium carbonate. Reported re-
sults are contradictory; Murray found a defi-
nite trend to higher reactivity but others have
reported little or no value.
In commercial calcining practice, verti-
cal and rotary kilns are the main types. The
rotary type is the more popular of the two in
the United States. Rotaries generally require
more fuel but less investment.
The main drawback to these established
types is in controlling the process to get max-
imum reactivity of product. Various designs
have been introduced to improve this situation.
An example is the Corson kiln, a nonrotating
drum that is shaken periodically to break up
the material and ensure uniform calcination.
Temperature and retention time are closely con-
trolled and particles are closely sized, so
that complete calcination can be obtained with-
out overheating. The resulting calcine has
very high reactivity.
Fluidized calcining is also used as a
means of getting uniform heat distribution and
avoiding hot spots. This is the type chosen
for the calcination cost estimates in this re-
port. It was selected over a vertical shaft,
rotary; or traveling grate kiln because it per-
4 Murray, J. A., Fischer; H. C., and Rolnick,
L. S. J. Am. Ceram. Soc. 37 (7), 323-328
(1954). -
73
-------
mits use of fines (-4 mesh) to produce calcine
without an unacceptably high loss of material.
Location of the calcination plant at a power
plant requires that an exceptionally high
yield of product be obtained from the raw lime-
stone; operation of a calciner by a diversi-
fied lime company permits use of undesirable
size fractions in other operations.
The system is designed for operation with
minus 1/4-inch feed stock. The kiln is a mul-
tistage unit with three beds for preheating,
one bed for calcining, and one bed for cooling.
Heat is supplied from combustion of coal be-
cause coal is readily available and is the low-
est cost fuel.
Coal is transferred from storage on a
belt conveyor to a feed hopper which supplies
the fuel to an air-swept pulverizer. This sys-
tem is similar to that used for combustion of
coal in a power plant boiler. The pulverized
coal and air mixture (primary air for combus-
tion) is ignited in burners mounted in the cal-
cining stage. Secondary air for combustion is
preheated as it passes countercurrently and
cools the solids discharged to the cooling
stage of the kiln. Sensible heat in the flue
gas heats the raw limestone as it passes count-
ercurrently in the preheating sections. Dust-
collection equipment is provided at the kiln
exhaust.
A belt conveyor is used to transfer lime-
stone from storage to a trash screen mounted
above the kiln feed hopper. A weigh belt feed-
er controls the rate of addition of stone to
the top preheat stage. Transfer between stages
is accomplished through use of gravity chutes
sealed from gas leakage by material on the low-
er bed. The calcined material discharges from
the cooling section of the kiln into a screw
conveyor; where further cooling occurs before
discharge into a storage bin. The bin will
accommodate sufficient storage for 12 hours'
operation to allow for minor delays in opera-
tion of the transfer belt, which conveys cal-
cined material to a 60-hour silo at the boiler.
From the silo, the calcined material is han-
dled in the same manner as discussed for lime-
stone injection.
The effect of plant size on investment
cost for a battery limits calcination plant is
given in Figure 11-4.
Hydration
The hydration of calcium oxide involves
about as many parameters as calcination of
limestone, and various types of equipment are
used. Dry hydration, the method assumed for
making hydrated lime for sulfur dioxide con-
trol, is more complicated than wet hydration
(slaking). Slaking could be used if slurry in-
jection were developed, but has not been con-
74
sidered in this study.
As in calcination, one of the problems in
hydration is to carry out the process in such
a way as to get optimum product properties.
Dolomitic calcine is the main problem, because
even if soft-burned it hydrates slowly and
some unhydrated material remains even after
long exposure to water. Several methods have
been developed for remedying this problem.
The Corson hydration system, for example, in-
volves hydration of the calcined particles
(usually 1/4 in. or larger) under pressure and
flashing into an expansion chamber. This pro-
cedure gives full hydration and a particle
size of about 0.25 micron.
Calcine made from high-calcium limestone
hydrates much more easily. The operation is
usually carried out at atmospheric pressure in
a vessel designed to mix the proper amount of
water with the calcine at controlled tempera-
ture. The product normally is air classified
to give the desired product size.
In making hydrated lime for sulfur diox-
ide sorption, it may be desirable to use a
pressure hydration system in order to get the
extremely small particle size obtainable by
this method. Further information on benefit
of small particle size is needed before a
choice can be made. For the cost estimates in
this report, the more widely known atmospheric
pressure system was assumed. Such a system
usually consists of a crusher to reduce lump
calcined lime to 1/2 inch or smaller, a screen
to separate the crushed material into selected
size fractions, storage bins for sized lime, a
conveyor system to transfer lime to a hydrator
feeder, a hydrator where controlled proportions
of lime and water are intimately mixed, a clas-
sifier for separation of coarse particles from
the hydrated product, and product storage
tanks.
The requirements for a hydration plant to
supply sorbent for boiler injection are some-
what different. Less stringent classification
is acceptable because a minor proportion of
coarse particles would not cause serious prob-
lems.
The system assumed in the present study
consists of:
1. A screw conveyor to transfer calcined
lime from the lime storage silo to the hy-
drator feed bin. A crusher is not required
since the calcined lime is small, minus 1/4
inch. It is assumed that the gradation is
acceptable for hydrator feed.
2. A volumetric feeder (screw type) to con-
trol the rate of addition of lime to the hy-
drator.
3. A Kritzer-type hydrator for operation
at atmospheric pressure. This type was cho-
sen because of its reported dependability
-------
en
0:
z
C
lLJ
X
IJ..
8
6
8000 HR. PER YEAR OPERATION
O~«,
«,S~
\,.. \ ~
4
2
HYDRATED L' ME
o
o 100 200 300 400 500 600
ANNUAL PLANT CAPAC ITV, THOUSAND TONS MATERIALS
Figure II-4.
Fixed Investment vs. Plant Capacity for Calcined Limestone
and Hydrated Lime (Battery Limits Cost)
and low maintenance costs.
A Kritzer hydrator is made up of a
stacked series of six cylindrical vessels
equipped with agitators. The calcined lime
is introduced in the top section and water
for hydration is added in a stack above the
top section to recover entrained solids
from the off-gas. The lime and water are
mixed and transported through the vessel by
the paddle-type agitator, and are dis-
charged to the second stage and progressive-
ly to the remaining stages in series.
4. Product conveyor (pneumatic). The hy-
drated product is discharged dry from the
final hydrator stage and is transferred by
a pneumatic conveyor to a storage bin sized
for 2 days' production.
5. From the storage bin, hydrated lime is
conveyed to the injection system surge tank
located at the boiler.
The effect of plant capacity on invest-
ment cost for a hydration plant is shown in
Figure II-4.
75
-------
APPENDIX III
EXPERIMENTAL WORK SUPPLEMENTING THE DESIGN STUDY
During the period June through November
1967) in the course of this study) experimen-
tal work was carried out at one of the 200-
megawatt boilers of the TVA Colbert Steam Plant
to obtain needed information on limestone in-
jection. Results of the tests are discussed in
this appendix. After the studies were made)
the results of other work) presented at the
Third Limestone Symposium (December 1967) and
summarized in Appendix I) became available.
Therefore the tests cover much of the same
ground as in Appendix I. They are presented
herein to make the coverage complete and to
supply supporting data for results reported by
other investigators.
Static Tests
Lumps (about 3/8 in.) of various lime ma-
terials were inserted into the boiler and held
there for 5 minutes. The lumps were then ex-
amined petrographically to determine degree of
reaction and nature of products. The findings
can be summarized as follows:
1. Essentially no sulfur dioxide was sorb-
ed below 1200° F.) even by calcined lime-
stone.
2. Calcined limestone and hydrated lime
sorbed carbon dioxide at 1200° F. and below.
3. Very little sulfur dioxide sorption oc-
curred in limestone lumps unless calcina-
tion was essentially complete.
4. Iron oxide impregnated in lumps of lime-
stone promoted both decarbonation and sul-
fur dioxide sorption.
5. There was occasional dead-burning in
samples exposed at 1900° F.
Injection Tests: Part Stoichiometric
A few tests were made) using equipment
readily available) in which about one-sixth
the stoichiometric amount of lime material was
injected by means of a sandblasting machine.
The type of equipment limited the injection
period to about 6 minutes. The objective was
to determine the degree of reaction by examin-
ing the lime after passage through the boiler.
Experimental difficulties) including erratic
injection rate and dilution of recovered sample
by fly ash) prevented gathering much signifi-
cant data. Typical analyses of the gross ash-
limestone mixture recovered are as follows:
76
Ca++, % S04--) %
Calcined limestone
Size
fraction)
mesh
Sorbent
utilization, a %
5.7
4.7
4.6
4.4
5.5
2.4
6.0
2.2
-325 b
Remainder
67
37
100
40
-325
Remainderb
Raw limestone
7.1
8.5
10.6
8.8
3.7
2.9
5.2
2.3
-325
Remainderb
-325
Remainderb
28
16
22
11
a Calculated by simultaneous equations from
analysis of ash collected just before injec-
tion period.
b About 50% -200 mesh.
Qualitative conclusions from these data
1. As indicated by S04:CaO ratio in
gross sample) calcined lime was more
tive than raw limestone.
2. In most of the tests with raw limestone
and dolomite) less than 30% of the injected
material was reacted.
3. Fine particles were reacted to a much
higher degree than coarser ones (all ma-
terials were ground to 70% -200 mesh before
injection). The degree of reaction indica-
ted that extremely fine particles would be
required for a high degree of sulfur dioxide
removal.
are:
the
effec-
Injection Tests: Full Stoichiometric
Full-scale tests were made to determine
whether limestone could be used effectively in
mixture with the coal. Although work by others
had indicated that some dead-burning may occur)
the data were not consistent.
Limestone was mixed with the coal in the
stoichiometric amount and the mix fed through
the various burners in such a way as to vary)
as much as possible within the limitations of
the equipment) the concentration of heat to
which the limestone was exposed and the reten-
tion time in the flame area. The boiler has
three horizontal rows of burners; the vertical
distance between the rows is about 6 feet.
-------
The following feeding arrangements were
used:
1. Through upper two rows with the bottom
one cut off.
2. Through the upper and lower rows with
the middle one off. This should give a
lower heat concentration than for 1.
3. Through all burners. This should give
the highest heat concentration.
Feeding into the boiler at a higher level,
to get a comparison with a point outside the
flame area, was not possible, of course, be-
cause no injection system for this was avail-
able.
Degree of sulfur dioxide removal was esti-
mated by (1) stack gas analysis, (2) solids
analysis, and (3) petrographic examination of
the solids. At least two samples were taken
in each test and averaged.
The results (Table III-I) checked fairly
well except for one anomalous gas analysis.
The degrees of sulfur dioxide removal ranged
from 21 to 27%. There was no clear indication
that injection pattern had any significant ef-
fect on sulfur dioxide sorption. The gas anal-
ysis from the test in which all burners were
fed limestone showed only 11% sulfur dioxide
removal (the anomalous value referred to),
whereas the solids analysis showed 21% reac-
tion of the limestone. Both were the lowest
values in their category. This mayor may not
indicate dead-burning with all burners going;
further tests are needed to clarify the point.
The 21 to 27% sulfur dioxide removal com-
pares favorably with results obtained by oth-
ers for injection at lower temperature, which
might indicate absence of dead-burning in the
current tests. However, there are so many
variables that it is difficult to compare tests
by different research organizations.
It may also be possible to evaluate dead-
burning by X-ray or other fine particle tech-
niques. Such tests are under way.
Other observations from the tests were:
1. The limestone ground as easily as the
coal and more of it ground to minus 200
mesh. Hardgrove grindability was 57 for the
limestone and 61 for the coal.
2. Slag accumulation was not excessive but
the stringy appearance was different from
normal.
3. The dust removal equipment (cyclone)
was not overloaded even though the solids
load was doubled.
Table III-I.
Test Summarya
Test No.
Mills on mix
Uni t load, mw.
Feed rate of mix, 1000 Ib./hr.
Average pulverizer current, amp. B mill
Theoretical air, %
Main steam temperature, of.
Main steam pressure, p.s.i.g.
Steam flow, 1000 Ib./hr.
Draft, inches of H20
ID fan inlet, A section
Air heater differential
Fly ash-CaO-S04' calculated ratio
Utilization of added CaO, %
Estimated by petrographic examination
Calculated from solids analysis
Calculated from gas analysis
1
ABCD
120
106
42
150
1065
1753
77
11.4
2.1
58-30-12
15 to 25
24
27
2
ABCDEF
150
131
36
132
1042
1753
97
12.4
2.7
55-33-12
15 to 25
21
11
3
ABEF
120
103
39
139
1008
1752
77
8.2
2.0
58-30-12
15 to 25
25
21
a See Figure III-l for burner arrangement.
77
-------
TO PROW
A a B MIL LS
MIDDLE ROW
caD MILLS
BOTTOM ROW
E a F MILLS
78
000000
808000
000080
Figure III-I. Colbert Unit 3 - Burner Arrangement
-------
APPENDIX N
COST ESTIMATES
Thermal Changes in Sorbent Injection: Effect
on Operating Cost
When limestone, calcined limestone, or hy-
drated lime is injected into a boiler for the
removal of sulfur dioxide from the exhaust
stack gases, a thermal change occurs due to
either the calcination of limestone or the de-
hydration of hydrated lime, plus the heat of
reaction of calcium oxide and sulfur dioxide
to form calcium sulfate. To compensate for
the resultant thermal effect and to maintain a
constant power output, a corresponding change
is required in the quantity of coal fed to the
boiler--with a resultant alteration in cost of
power generation. This change in power plant
operating cost must be charged to dry sorbent
injection.
The bases and methods of determining this
cost are shown below:
Bases
Plant size, 200 mw.
3.5% sulfur in. coal
200% stoichiometric calcium oxide addition
33% calcium oxide reacted with sulfur dioxide
12,000 B.t.u./lb. coal burned
Exit stack gas temperature, 300° F.
Air and material injected at 50° F.
2 lb. air injected/13 lb. of sorbent
Moisture content
Limestone, 10%
Calcined lime, none
Hydrated lime, 1%
1.
Calculations
Limestone (139,000 tons/yr.)
Heat required
Calcination
(74,000 tons x 2.8 MM
B . t . u . / to n C aO )
Limestone sensible heat
(139,000 tons x 2000 Ib./ton
x 0.24 sp. ht. x 250°)
Moisture sensible heat
(13,900 tons x 2000 Ib./ton
x 1.0 x 250°)
Moisture latent heat
(13,900 tons x 2000 Ib .jton
x 950 B.t.u./lb.)
Air sensible heat
(20,000 tons x 2000 Ib./ton
x 0.26 x 250°)
Total heat used
(equivalent to 3.51 MM
B.t.u./ton of caO)
Millions of
B.t.u./yr.
207,200
16,680
6 , 950
26,410
2,600
259,840
Heat
Heat of reaction
CaO + S02 +
1/202 ~ CaS04
generated
CaS04 formed
Total heat generated
3.00 MM B.t .u./
ton CaS04
59,350 tons/yr.
178,050 MM
B.t.u./yr.
81,790 MM
B.t.u.jyr.
Resultant thermal deficit
Makeup coal required
(based on 12,000 B.t.u./
lb. coal)
3,400 tons/yr.
From analysis of TVA and private power
plant operating costs directly chargeable
to quantity of coal fed, the following
charges are assumed.
Raw material
Operating labor, maintenance,
utilities, and overhead
Annual capital charge, at 11%
of coal-related investment,
of $6.62/ton coal burned
Total cost
$/ton coal
4.20
1.42
Q..:12
b:35
Total deficit cost of thermal
effect: 3400 tons x $6.35 = $21,600
2.
(74,000 tons/yr.)
Millions of
B.t.u./yr.
8,840
1,380
10 ,220
Calcined Limestone
Heat required
Lime sensible heat
Air sensible heat
Total heat used
Heat generated
178,050 MM B.t.u./yr.
167,830 MM B.t.u./yr.
Same as in 1
Resultant thermal
credit
Decrease in coal
requirement
7,000 tons/yr.
Total credit cost
to thermal effect:
7,000 tons x $6.35 = $44,450
3.
Hydrated Lime (98,600
tons/yr. )
Millions of
B.t.u./yr.
74,000
12,325
2,366
1,972
90 ,065
Heat required
Dehydration heat
Hydrate sensible heat
Free water sensible heat
Air sensible heat
Total heat used
79
-------
Heat generated
Same as in ! and ~ 178,050 MM B.t.u./yr.
Resultant thermal
credi t
Decrease in coal
requirement
87,985 MM B.t.u./yr.
3,666 tons/yr.
Total credit cost
to thermal effect:
3,666 tons x $6.35 = $23,280
Cost and Benefit of Fine Grinding
With adequate data, it should be possible
to optimize the grinding system by calculating
the particle size that gives the best balance
between cost of grinding and sorption efficien-
cy. Such a calculation would allow selection
of the best combination of particle size and
amount of sorbent to give the lowest overall
cost for a given degree of sulfur dioxide re-
moval.
The cost of grinding has been estimated
by calculating the cost of grinding to 70% mi-
nus 200 mesh based on TVA coal-grinding expe-
rience. The cost of grinding to smaller sizes
was then estimated from data on limestone
grinding reported by Work and Stern.l
Based on TVA power plant experience, the
cost of grinding coal to 70% minus 200 mesh in
a 17.5-ton-per-hour bowl mill is summarized
below:
Labor and overhead
Power
Maintenance
Capital charge at 11%
of investmenta
Cost of
grinding,
cents/ton
3.0
5.2
11.4
12.4
32.0
a A coal pulverizer capable of
grinding 140,000 tons per year
for a 200-megawatt power plant
is estimated to cost $156,000.
Assuming that the rate of grinding lime-
stone will equal that of coal,2 the above-
stated figures can be used as base cost of
grinding limestone to 70% minus 200 mesh. To
grind limestone from this size to 99% minus
325 mesh, for example, Work and Stern show a
capacity decrease of 65% and a power increase
of 250% per ton of limestone ground with the
following resultant cost effect.
1 Work, L. T., and Stern, A. L. In Chemical
Engineers' Handbook (R. H. Perry, C. H. Chil-
ton, and S. D. Kirkpatrick, eds.) pp. 8-47
McGraw-Hill (1963). "
2 ~is a~pear~ reasonable, since the published
work lndex for coal (11.37) is almost the
same as that for limestone (11.2).
80
Cents/
ton
Power (5.2 cents x 2.5)
'::::~ema~~i~)nce, capital
Total cost of grinding to 99%
-325 mesh
Incremental cost of grinding
from 70% -200 mesh to 99%
-325 mesh
13.0
76.6
89.6
57.6
Costs for other particle sizes were also
estimated from capacity and power figures given
by Work and Stern. The results are given in
the form of a curve in Figure 9 (main text).
The costs appear conservative because Work and
Sterns's data were obtained in a mill about
half the size required for the 200-megawatt
unit. The larger mill should give lower grind-
ing cost.
Not only mill size but also mill type,
variation in limestone hardness and moisture
content, and probably other factors all affect
cost of grinding. Therefore a generally appli-
cable relationship is not easy to develop.
Probably the nearest thing to this is the for-
mula proposed by Bond.3
W = 10 Wi - 10 Wi
pO.5 FO.5
(1 )
where W = kw.-hr./ton of material ground
Wi = a work index for a given mate-
rial (based on experience)
F = particle size (in microns) of
the feed material, in terms of
80% minus the "F" size
P = particle size of product on the
same basis
The work index for limestone, based on measure-
ments with 181 samples, averaged 11.2. If we
assume that overall cost of grinding is propor-
tional to power requirement, which is approxi-
mately correct, and that the ratio between the
two is the same as in the calculations for Fig-
ure 9 (average of 6.6), then a curve (Fig. IV-l)
can be drawn based on Bond's formula. The
equation for the curve is of the form:
x = ayn - b (2)
However, for the smaller particle size ranges
the second constant has little effect and the
curve can be represented by:
c = 2.1d-o.5
where c = cost/ton
d = particle
(3 )
diameter
3 Bond, F. C.
(1967) .
Rock Products 70 (3), 132-133
-------
10.00
1.00
5.00
z
~ 3.00
......
.(f)-
..
~ 2.00
o
t; 1.50
w
:E
..J 1.00
~
z
~ 0.70
0::
C)
IJ.. 0.50
o
..... 0.40
f/)
8 0.30
0.20
0.10
I
1.5
2
2.5 3 4 5 6 1 8 9 10 20 30 40 50 60
PARTICLE SIZE, MICRONS (80 PER CENT OF
MATERIAL SMALLER THAN GIVEN MICRON SIZE)
80 100
Figure IV -1.
Grinding Cost of Limestone Calculated from Bond's Equation
The two curves, one based on actual data
and the other on a semitheoretical relation-
ship, differ mainly in that the Bond formula
gives a higher cost in the 5-micron range--
which makes it conservative. It has the advan-
tage of being representable by an equation
whereas the other is not easily amenable to
curve fitting.
The benefit of fine grinding is more dif-
ficult to estimate than the cost because of
scarcity of data. One approach to determining
the relative effect of particle size on sorp-
tion efficiency is the mathematical one of as-
suming that a calcium sulfate shell of a cer-
tain thickness is formed--the thickness depend-
ing on sorbent reactivity, temperature of in-
jection, and retention time in the boiler--and
that the thickness of this shell does not vary
with particle size. This is an idealized sit-
uation that is not likely to hold but the con-
cept helps in evaluating the benefit of fine
particle size.
If such a shell of nonvarying thickness is
formed, it can be shown that the following re-
81
-------
lationship holds--assuming that the particles
are spheres, the shell is completely reacted,
and the portion inside the shell is completely
unreacted.
x = a3 - 3a2r + 3ar2 x 100
r3
(4 )
where x = percent reaction of the lime-
stone when a stoichiometric
amount is injected
a = shell thickness in microns
r = particle radius in microns
From consideration of the degrees of sorp-
tion that have been obtained in actual tests,
it can be seen that the shell must be relative-
ly thin in the medium size range of, say, 100
to 10 microns. This being the case, the a3
and 3a2r terms in Equation 1 become small, and,
as an order of magnitude, the percent utiliza-
tion becomes inversely proportional, approxi-
mately, to the first power of the radius.
This would be quite desirable if true, because
Bond's formula indicates that the cost of
grinding is inversely proportional to the 0.5
power of the radius. Hence the percent utili-
zation would increase faster than grinding
cost as particle size decreased.
Test results, however; have shown that
the benefit of fine grinding is much less than
the "unvarying thickness of shell" theory indi-
cates. Apparently the only data available in
a sufficient amount and form to allow a deter-
mination of the relationship between particle
size and degree of reaction are those of
Ishihara and Tanaka. Each plotted the data on
a log log basis and obtained a straight line,
indicating a relationship of the type:
x = brn
(5 )
This is quite different from Equation 4, and
indicates that the shell effect, if any, is
not based on uniform shell thickness.
Ishihara's datal can be represented by
the equation:
x = 0 .272r-o .33 x 100 (6)
over the range of 3 to 40 microns. (The basis
for the particle sizes is not clear from Ishi-
hara's report; it is assumed that the value
given is for the particle size at which half
the weight of material is larger and half
smaller. )
Tanaka did not do as well; his data fol-
low the relationship:
x = 0.222r-O.137 x 100
(7)
The departure of these equations from the
shell theory can be illustrated by Ishihara's
1 See Appendix I.
82
data. At 40-micron diameter, the 10% utiliza-
tion obtained represents a theoretical shell
thickness of about 0.75 micron. At 4-micron
diameter, however, a thickness of only 0.143
micron is indicated for the 22% sorption; with
a 0.75-micron shell, degree of reaction would
have been about 75%.
The reason for the poor performance at
small particle size is not clear. Zentgraf al-
so noted this effect in tests with hydrated
lime and attributed it, in part, to particle
agglomeration and to reduced mass transfer be-
cause of lower "drag" in the stack gas as com-
pared with larger particles. In addition, the
greater tendency of fines to dead-burn may have
been a factor, as each set of tests was at the
same temperature, and decreasing sulfur diox-
ide driving force at higher degrees of utiliza-
tion may have affected the sorption.
Even if sorption does not increase as
fast as might be expected when particle size
is decreased, there is still an optimum parti-
cle size, which can be obtained by the rela-
tionship:
L + G
C=-
U
(8)
where C = cost of limestone actually uti-
lized for sulfur dioxide sorp-
tion, per ton
L = cost of delivered limestone,
per ton
G = cost of grinding, per ton for a
given particle size, d
U = percent utilization (divided by
100) for the given particle size,
d
In terms of r; and using Bond's equation,
this becomes:
L + ar-n
br-m
c =
(9)
If m is smaller than n, as the previous dis-
cussion has shown, then the resulting curve is
a hyperbola with a low, or optimum, point.
Substituting Equations 3 and 6 (Ishihara's
data) gives:
L + 1.48r-o .5
c = 0.272r-O.33
Tanaka's data (Equation 7) gives the same type
of curve but with a different optimum point be-
cause of the smaller effect he found for parti-
cle size.
Curves from these equations are given in
Figure 10 (see main text), assuming $2 per ton
delivered limestone cost. In plotting the
curves it was necessary to estimate a conver-
sion factor from Bond's "80% minus given size"
to Ishihara and Tanaka's "mean diameter." This
(10 )
-------
was done from a typical size distribution of
finely ground limestone. With all these con-
versions and assumptions, it is obvious that
the curves do not have great quantitative sig-
nificance. However, they show the general re-
lationships involved and afford a method for
optimizing grinding when adequate data become
available.
Detailed Estimates
Table IV -I.
Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Limestone Injection Process--Case Ia
Annual quantity
Unit cost, $
Total
annual
cost, $
Cost/ton
of coal
burned, $
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent injection
Precipitator
Maintenance
Labor and material
Analyses
39.6 M ton
2.05/ton
3.00/hr.
1,500,000 kw.-hr.
379,000 kw. -hr.
0.003/kw. -hr.
0.003/kw. -hr.
81,200 0.135
32,000 0.020
4,500 0.007
1,100 0.002
40,000 0.067
7,000 0.032
64,600 0.108
145,800 0.243
4,000 hr.
1,000 hr.
1.00
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
Occupancy
Interest, 6% of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of fixed investment
Taxes and insurance, 2% of fixed investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect costs
Total limestone injection cost
37,100 0.062
46,000 0.077
18,400 0.031
32,900 0.021
1,200 O. 002
U5 ,600 0.193
261,400 0.436
6,100 0.010
261,500 0.446
Thermal effect of limestone injection on
operating cost of power plant
Total annual operating cost
a Basis:
200-mw. system
Coal burned, 600,000 ton/yr.
Sulfur in coal, 2.0%
Electrostatic precipitator operation
North Alabama plant location
Limestone added, 100% of stoichiometric
Limestone injected, ground to 70% -200 mesh
Coal, 0.75 Ib./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $920,000 fixed; $8,100 working
83
-------
Table IV -II .
Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Limestone Injection Process--Case IIa
47,000 0.078
58,000 0.097
23,200 0.039
16,600 0.028
1,400 0.002
146,200 0.244
372,000 0.620
10,800 0.018
382,800 0.638
Annual quantity
Unit cost, $
Total
annual
cost, $
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent i~iection
Precipitator
Maintenance
Labor and material
Analyses
69.5 M ton 2.05/ton 142,500
4,667 lIT. 3.00/lIT. 14,000
2,284,000 kw.-hr. 0.003/kw. -lIT. 6,800
660,500 kw. -lIT. o. 003!kw. -lIT. 2,000
1,500 lIT.
7.00
50,000
10,500
Subtotal conversion costs
Subtotal direct costs
83,300
225,800
Indirect Costs
Occupancy
Interest, 6% of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of fixed investment
Taxes and insurance, 2% of fixed investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect. costs
Total limestone injection cost
Thermal effect of limestone injection on
operating cost of power plant
Total annual operating cost
cost/ton
of coal
burned, $
0.238
0.023
0.011
0.003
0.084
0.017
0.138
0.376
a Basis:
200-mw. system
Coal burned, 600,000 tons/yr.
Sulfur in coal, 3.5%
Electrostatic precipitator operation
North Alabama plant location
Limestone added, 100% of stoichiometric
Limestone injected, ground to 70% -200 mesh
Coal, 0.75 Ib./kw.-lIT.
Unit on-stream time, 8000 hr./yr.
Capital investment, $1,160,000 fixed; $14,200 working
84
-------
Table IV-III. Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Limestone Injection Process--Case IlIa
Direct Costs
Annual quantity. Unit cost, $
Delivered raw material
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent injection
Precipitator
Maintenance
Labor and material
Analyses
Subtotal conversion costs
79.2 M ton
4,933 hr.
2,573,000 kw.-hr.
750,400 kw.-hr.
1,800 hr.
Subtotal direct costs
Indirect Costs
Occupancy
Interest, 6% of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of fixed investment
Taxes and insurance, 2% of fixed investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect costs
Total limestone injection cost
Thermal effect of limestone injection on
operating cost of power plant
Total annual operating cost
2.05/ton
3.00/hr.
0.003!kw. -hr.
0.003/kw. -hr.
7.00
Total
annual
cost, $
Cost/ton
of coal
burned, $
162,400 0.270
14,800 0.025
7,700 0.013
2,200 o. 004
53,900 0.090
12,600 0.021
91,200 0.153
253,600 0.423
50,200 0.084
62,000 0.104
24,800 0.041
18,200 0.030
1,500 o. 002
156,700 0.261
410,300 0.684
12 ,200 0.020
422,500 0.704
a Basis:
200-mw. system
Coal burned, 600,000 tons/yr.
Sulfur in coal, 2.0%
Electrostatic precipitator operation
North Alabama plant location
Limestone added, 200% of stoichiometric
Limestone injected, ground to 70~ -200 mesh
Coal, 0.75 Ib./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $1,240,000 fixed; $16,200 working
85
-------
Table rv -rv .
Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Limestone Injection Process--Case rva
Annual quantity
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent i~iection
Precipitator
Maintenance
Labor and material
Analyses
139.0 M ton
6,267 hr.
4,000,000 kw.-hr.
1,321,000 kw.-hr.
2,960 hr.
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
Occupancy
Interest, 6~ of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of fixed investment
Taxes and insurance, 2% of fixed investment
Overhead
Plant, 20% of conversion casts
Administrative, 10% of operating labor
Subtotal indirect costs
Total limestone injection cost
Thermal effect of limestone injection on
operating cost of power plant
Total annual operating cost
Unit cost, $
2.05/ton
3.00/hr.
O. 003/kw. -hr.
O. 003/kw. -hr.
7.00
Total
annual
cost, $
cost/ton
of coal
burned, $
285,000 0.475
18,800 0.031
12,000 0.020
4,000 0.007
69,500 0.116
20,700 0.034
125,000 0.208
410,000 0.683
63,100 0.105
77 ,500 0.129
31,000 0.052
25,000 0.042
1,900 0.003
193 ,500 0.331
608 , 500 1. 014
21,600 0.036
630,100 1. 050
a Basis:
200-mw. system
Coal burned, 600,000 tons/yr.
Sulfur in coal, 3.5%
Electrostatic precipitator operation
North Alabama plant location
Limestone added, 200% of stoichiometric
Limestone injected, ground to 70% -200 mesh
Coal, 0.75 lb./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $1,550,000 fixed; $28,500 working
86
-------
Table IV -V.
Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Limestone Injection Process--Case Va
Annual quantity
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent injection
Precipitator
Maintenance
Labor and material
Analyses
695.0 M ton
12,000 hr.
12,000,000 kw.-hr.
6,605,000 kw.-hr.
10,800 hr.
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
Occupancy
Interest, 6% of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of fixed investment
Taxes and insurance, 2% of fixed investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect costs
Total limestone injection cost
Thermal effect of limestone injection on
operating cost of power plant
Total annual operating cost
Unit cost, $
2.05/ton
3.00/hr.
0.003!kw. -hr.
0.003/kw.-hr.
7.00
Tot 9.1
annual
east, $
Cost/ton
of coal
burned, $
1,425,000 0.475
36,000 0.012
36,000 0.012
19,800 0.007
347,500 0.116
75,600 0.025
514,900 0.172
1,939,900 0.647
163,700 0.054
197 ,500 o. off)
79,000 0.026
103,000 0.034
3,600 0.001
546,800 0.180
2,486,700 0.827
108,000 0.036
2,594,700 0.863
a Basis:
1000-row. system
Coal burned, 3,000,000 ton/yr.
Sulfur in coal, 3.5%
Electrostatic precipitator operation
North Alabama plant location
Limestone added, 200% of stoichiometric
Ground limestone injected, 70% -200 mesh
Coal, 0.75 Ib./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $3,950,000 fixed; $142,500 working
87
-------
Table IV-VI. Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Calcined Limestone Injection Process--Case Va
Annual Quantity
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent injection
Precipitator
Coal (calcining)
Maintenance
Labor and material
Analyses
730.0 M ton
62,000 hr.
20,803,000 kw.-hr.
6,605,000 kw.-hr.
76,500 tons
12,000 hr.
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
Occupancy
Interest, 6% of borrowed capital
(assume 66% borrowed)
Depreciation, 5% of plant investment
Taxes and insurance, 2% of plant investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect costs
Total calcined limestone injection cost
Thermal effect of calcined limestone
operating cost of power plant
Total annual operating cost
injection on
Unit cost. $
2.05/ton
3.00/hr.
0.003/kw. -hr.
o. 003/kw. -hr.
4.26/ton
7.00
Total
annual$
cost.
Cost/ton
of coal
burned, $
1,500,000 0.500
186,000 0.062
62,400 0.021
19,800 0.007
321,300 0.107
587,500 0.195
84,000 0.028
1,261,000 0.420
2,761,000 0.920
410,000 0.136
491 ,500 0.166
199,000 0.067
252,200 0.084
18 ,600 0.006
1,377,300 0.459
4,138,300 1.379
(221. 600) (0.074 )
3,916,700 1.305
a Basis:
1000-mw. system
Coal burned, 3,000,000 tons/yr.
Sulfur in coal, 3.5%
Electrostatic precipitator operation
North Alabama plant location
Calcined limestone added, 200% of stoichiometric
Ground calcined limestone injected, 70% -200 mesh
Coal, 0.75 Ib./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $9,950,000 fixed; $300,000 working
88
-------
Table IV-VII. Annual Operating Cost for Sulfur Dioxide Removal from Power Plant
Stack Gas: Dry Hydrated Lime Injection Process--Case Va
Direct Costs
Delivered raw materials
Limestone, 95% C03
Conversion costs
Operating labor
Utilities
Sorbent i~iection
Precipitator
Water
Coal (calcining)
Maintenance
Labor and material
Analyses
Annual quantity
Unit cost, $
Total
annual
cost, $
766.0 M ton 2.05/ton 1,570,300
95,333 hr. 3.00/hr. 286,000
23,309,000 kw.-hr. 0.003!kw. -hr. 69,200
6,605,000 kw.-hr. O. 003/kw. -hr. 19,800
75,000 M gal. O.d+/M gal. 3,000
76,500 tons 4.20/ton 321,300
13,000 hr.
710,500
91,000
1,500,800
7.00
Subtotal conversion costs
Subtotal direct costs
Indirect Costs
3,071,100
Occupancy
Interest, 6~ of borrowed capital
(assume 66% borrowed)
Depreciation, 5~ of plant investment
Taxes ~nd insurance, 2~ of plant investment
Overhead
Plant, 20% of conversion costs
Administrative, 10% of operating labor
Subtotal indirect costs
Cost/ton
of coal
burned, $
0.523
0.095
0.023
0.007
0.001
0.107
0.237
0.031
0.501
1. 024
493 ,000 0.166
600,000 0.200
240,000 0.080
300,300 0.100
28,600 0.009
1,666,900 0.555
4,738,000 1. 579
(115,900) (0.039)
4,622,100 1. 540
Total hydrated lime injection cost
Thermal effect of hydrated lime injection on
operating cost of power plant
Total annual operating cost
a Basis:
1000-row. system
Coal burned, 3,000,000 tons/yr.
Sulfur in coal, 3.5%
Electrostatic precipitator operation
North Alabama plant location
Hydrated lime added, 200% of stoichiometric
Coal, 0.75 Ib./kw.-hr.
Unit on-stream time, 8000 hr./yr.
Capital investment, $12,000,000 fixed; $450,000 working
89
-------
'8
/ I!
~. (\ / ~-~) .~-'(-'J
i 4.. [' -- --_h~~ +-+-]-
.' ..~~r +
I - - W' "M~~lON" ''''IlAG4
UM~'TON~ N~V"V~ ( ~/,a ("" HI. ~A)
HOH~" ($""- c.v<)
I --
i - - - U_n__- I~Y
>-cJ
~
~
t:J
H
~
<:
-------
//
(--
I
----=:::::::::;:..--- -
~~I~l
I
-- -~---r'/
"" -
--~ ~
,,-- , . -
_---~~;1; -
'----- - ------- iL----:--..---:
'---. --iJ<------
.- . <"-'-...
~--
-~
Figure V-2. Colbert Steam Plant
Removal of Sulfur Dioxide from Power Plant Stack Gas
Limestone Injection Process: Perspective View
\.0
t--'
------- |